[00:00:04] Welcome back to another episode of Tech

[00:00:06] Unhinged, where tech gets human. I’m

[00:00:07] your host, Rabbi Davidid, and today we

[00:00:09] are joined by someone who has spent over

[00:00:11] 15 years helping consulting firms stop

[00:00:13] being invisible and start winning better

[00:00:15] work. Florian Henrik is the founder of

[00:00:17] client friendly and a partner at the

[00:00:19] visible authority both focused on

[00:00:21] helping midsize consulting firms. He

[00:00:23] works with firms ranging from 3 to 10

[00:00:25] million in revenue. Before that, he held

[00:00:27] senior in-house roles at Deoid and

[00:00:29] Essenture where he helped build and

[00:00:31] scale a multi-billion dollar business

[00:00:33] unit driving towards 25% yearon-year

[00:00:37] growth targets. He also brings deep

[00:00:39] experience from senior roles at global

[00:00:41] communications agencies BCW global and

[00:00:44] Levis communications advising

[00:00:46] professional services and technology

[00:00:47] firms across international markets.

[00:00:49] Welcome to the show, Florian.

[00:00:51] >> Thank you for having me.

[00:00:53] >> Well, um we’re so glad to have you with

[00:00:55] us today. We’ll start off with some ice

[00:00:57] breakers before we dive into the topic

[00:00:59] is how AI is reshaping tech

[00:01:01] consultancies in 2026. So Floren, given

[00:01:05] the enormous scope of your consulting

[00:01:07] experience and leadership services, if

[00:01:09] you had to describe consulting today in

[00:01:11] one word, would it still be expertise or

[00:01:14] is it becoming automation?

[00:01:16] >> It’s expertise and it will remain

[00:01:17] expertise forever. And my pointed

[00:01:19] response there was if software in itself

[00:01:21] would solve any problems, consulting

[00:01:23] would have been dead decades ago. But it

[00:01:26] turns out that it continues to be true.

[00:01:27] You need both the expertise and the

[00:01:29] people side of things and the computers

[00:01:31] and the software to really get to a

[00:01:33] solution. And I think that’s what the

[00:01:35] game continues to be. Right.

[00:01:37] >> Well, I think your your this answer kind

[00:01:39] of nullifies my next question because it

[00:01:41] was that if I’m

[00:01:42] >> sorry for that. If if there was one

[00:01:44] thing you would say that still feels

[00:01:46] uniquely human in the world of

[00:01:47] automation in consulting, what would it

[00:01:50] be?

[00:01:51] >> Well, I think there’s lots of talk about

[00:01:52] creativity and judgment and and I think

[00:01:55] all of those remain true. I would add to

[00:01:58] that to those notions that if on a

[00:02:00] certain philosophical level, I mean what

[00:02:02] is business or what is a business is

[00:02:04] people working with people and then

[00:02:06] machines come into this at a varying

[00:02:07] degree. Tools or machines have been for

[00:02:10] thousands of years. So it’s a mix of the

[00:02:12] two but the moment people side of the

[00:02:15] equation disappears I think I might be

[00:02:18] wrong but I think there’s no business

[00:02:20] anymore. So that that that stuff will

[00:02:22] remain relevant. I have said elsewhere

[00:02:24] that I also think the the people centric

[00:02:26] aspects of consulting might actually

[00:02:28] become ever more important if if you

[00:02:30] think about what AI does to a certain

[00:02:32] extent or or maybe to be more careful

[00:02:35] claims to do because there’s a

[00:02:37] difference between what the tools claim

[00:02:38] to be able to do and then what they are

[00:02:40] actually able to do today. What they

[00:02:42] claim to do is take out a lot of the

[00:02:44] toil work in consulting, right?

[00:02:45] Crunching the numbers, building the

[00:02:48] model, running the simulation, writing

[00:02:50] the plan, like drafting like a

[00:02:52] deliverable or whatever. And that’s

[00:02:54] fine, but that doesn’t take away the

[00:02:56] need for human judgment, for

[00:02:58] communicating about the deliverables,

[00:02:59] for working with people, addressing

[00:03:01] concerns, all these types of things.

[00:03:03] Like, let’s call that the change aspect

[00:03:05] or the implementation aspect. In fact,

[00:03:07] what I believe is what it does, it

[00:03:09] heightens the need for those because now

[00:03:11] that the computer can do many of the

[00:03:13] preparatory tasks, there will be more of

[00:03:15] that done in shorter time, which means

[00:03:17] the demand for people who work on the

[00:03:19] finalizing the decision and then making

[00:03:21] the change happen side of things will

[00:03:24] probably go up. I think I I might

[00:03:26] mispronounce the name. It’s I think it’s

[00:03:28] called the Javans paradox that describes

[00:03:30] just that if processes get more

[00:03:32] efficient that efficiency usually gets

[00:03:34] reinvested not in savings but in

[00:03:36] increased consumption of the now more

[00:03:38] efficient process. So with AI coming in

[00:03:40] I would expect the need for consulting

[00:03:41] to go up. Does that make sense? Um I’m

[00:03:44] not saying the jobs stay the same. I’m

[00:03:46] not I’m not saying the jobs will stay

[00:03:47] the same. I’m not saying the cost will

[00:03:48] stay the same. There are there will

[00:03:49] there’s massive change happening already

[00:03:51] in the market for sure. But my

[00:03:52] prediction is that now that some of the

[00:03:55] labor intensity and maybe some of the

[00:03:56] costs come down usually demand goes up.

[00:03:59] >> So now if when we talk about the

[00:04:00] breaking point here and how AI is

[00:04:02] actually you know disrupting we see that

[00:04:05] there is this growing narrative

[00:04:06] including from Forbes tech council that

[00:04:09] AI is fundamentally breaking the

[00:04:11] traditional consulting model as you

[00:04:13] hinted upon you know in your previous

[00:04:14] answer. So what exactly is breaking and

[00:04:17] why now?

[00:04:18] >> I wish I knew that. I’m working hard to

[00:04:21] figure that out every day. This is it’s

[00:04:22] a complicated story and I like that you

[00:04:24] use the word the narrative is there

[00:04:25] because the first thing you can see

[00:04:27] immediately is there’s hype around AI

[00:04:29] solutions in the market and that hype

[00:04:31] has an immediate impact. Techminded

[00:04:33] critical people sometimes struggle to

[00:04:35] grasp because they might say well but

[00:04:37] the AI cannot do this or that and they

[00:04:39] are right uh but that’s not the point.

[00:04:41] The point is that some business sponsor

[00:04:44] somewhere believes that maybe the

[00:04:46] technology can right and so their

[00:04:48] willingness to pay for the alternative

[00:04:50] and that’s the first impact is eroding

[00:04:53] rather quickly and you’ll have read the

[00:04:54] headlines right I mean I think one of

[00:04:56] the big four consultancies themselves

[00:04:58] demanded a discount for their own

[00:05:00] auditor claiming that that auditor was

[00:05:02] now doing parts of the audit process

[00:05:04] with AI so why isn’t this cheaper so the

[00:05:07] belief in AI enabling a utopian future

[00:05:11] future of doing everything through the

[00:05:12] computer is already changing the buying

[00:05:15] behavior. That is impact number one.

[00:05:17] Impact number two then is that because

[00:05:20] of those expectations, there is now a

[00:05:22] huge expectation on consulting leaders

[00:05:24] to reinvent their workflows and their

[00:05:26] practices with AI. That expectation is

[00:05:28] there also and it can go two ways.

[00:05:31] Either it works and the technology can

[00:05:33] take some of the work away. Well, now

[00:05:36] you have to restructure headcounts and

[00:05:37] teams and stuff because computer can do

[00:05:39] certain things that consultants used to

[00:05:41] do in the past. That is however the good

[00:05:43] case. What we hear a lot about behind

[00:05:46] closed doors is the opposite case where

[00:05:48] the technology doesn’t do good enough

[00:05:50] work to meet the expectations which

[00:05:51] leaves consulting leader in in a very

[00:05:53] tricky double bind position. We we have

[00:05:55] many senior managers, directors,

[00:05:57] managing directors talk to us about

[00:05:58] this. The situation being there’s a

[00:06:00] hiring freeze or people have been let go

[00:06:02] even because hey you should be able to

[00:06:04] do this with AI but the AI can’t do it.

[00:06:06] So now you have more work that goes to

[00:06:08] more senior people while you’re dealing

[00:06:10] with the technology that for reasons

[00:06:13] that could be different isn’t exactly

[00:06:15] delivering against the job specification

[00:06:16] and that is that is a severe problem on

[00:06:18] on teams everywhere right now. You could

[00:06:20] hatch some of that with just accepting

[00:06:22] lower quality, right? But again, we’ve

[00:06:24] read in the press that that is probably

[00:06:26] not a feasible solution because again,

[00:06:28] as in the case of a brick four, you

[00:06:29] might be dragged in front of a court and

[00:06:31] being sued because the reports you

[00:06:32] delivered were full of mistakes and AI

[00:06:35] made and you are being held accountable

[00:06:37] rightfully so for that. So what breaks I

[00:06:40] think is these three things. So there’s

[00:06:41] changes to willingness to pay. There is

[00:06:43] what what we have called the AI double

[00:06:45] squeeze. So this mismatch of expectation

[00:06:47] versus what you can really do with the

[00:06:48] tooling. And then there are areas where

[00:06:50] the technology is absolutely phenomenal

[00:06:52] in automating stuff and you got to be

[00:06:54] smart about what that means for your for

[00:06:56] how you do work. Put all this together

[00:06:59] and what happens is the business model

[00:07:01] the specific business model variant of

[00:07:03] consulting that is built on leverage of

[00:07:06] capacity. So I am one very experienced

[00:07:09] person that knows their stuff and I

[00:07:11] service many clients and many packages

[00:07:13] of work because I have more junior

[00:07:15] people underneath me. I leverage their

[00:07:17] capacity. That type of consulting is

[00:07:19] under pressure because at least some

[00:07:21] parts of that capacity can come out of a

[00:07:23] machine today and very cheaply because

[00:07:26] because it’s heavily subsidized, right?

[00:07:27] No one’s paying the real cost for this

[00:07:29] AI stuff, which is another like what’s

[00:07:31] breaking now. We talked about what might

[00:07:33] break in the future. Well, the economics

[00:07:35] are not exactly behind those large

[00:07:36] language models. So, if they change

[00:07:38] their prices, all my answers will

[00:07:40] probably have to change again. But but

[00:07:42] for now it’s alluringly cheap to

[00:07:44] automate the bits and pieces of your

[00:07:46] work that can be automated with certain

[00:07:48] degrees of reliability. So that’s

[00:07:49] happening as well.

[00:07:50] >> We see that how the financial time

[00:07:52] reports firms are already restructuring

[00:07:54] around AI not just on the outside but

[00:07:57] internally as well and we are also

[00:07:58] seeing a genuine operational change you

[00:08:01] know within certain firms. So, you know,

[00:08:03] when we when we have a deeper look at

[00:08:05] it, do do you think that most of it is

[00:08:08] still just, you know, rebranding and

[00:08:10] marketing and there is like no ground

[00:08:12] reality to it?

[00:08:13] >> This is what makes all these questions

[00:08:14] so difficult to answer because there is

[00:08:17] some ground reality to it. Like if you

[00:08:19] have I’m not a software engineer, but I

[00:08:22] some code around the edges of business

[00:08:24] processes for for funsies. If you work

[00:08:26] with clo it seems amazing what it can

[00:08:28] do. There’s certainly parts in just

[00:08:31] toile work in the office. We are

[00:08:33] recording this call right now. So, so

[00:08:35] writing summaries like these types of

[00:08:36] things. There is genuine value there.

[00:08:39] That’s one part. The other part is does

[00:08:42] that is that enough right to to make up

[00:08:46] for the insane cost of it should be

[00:08:48] clear, right? If you talk about AI right

[00:08:50] now, I’m presuming we’re predominantly

[00:08:52] talking about the generative stuff that

[00:08:53] comes to us in the forms of large

[00:08:55] language models and how they can be

[00:08:56] deployed in in in knowledge work. It’s

[00:08:58] insanely expensive, incredibly so. To

[00:09:01] the point where I don’t know if you read

[00:09:04] those numbers, $600 billion in in capex

[00:09:06] investments. And is that going to ever

[00:09:09] be is that ever going to create a return

[00:09:11] by helping people write email better to

[00:09:14] to be a bit provocative? That remains to

[00:09:16] be seen. and and specifically with

[00:09:18] regards to software development which

[00:09:20] that seems to be the discipline where it

[00:09:22] could have the gravest impact because

[00:09:24] there’s a constraint logic you can check

[00:09:26] against right it is code code writing is

[00:09:29] a form of text generation which is what

[00:09:31] those models are designed to do so so I

[00:09:32] can see that it can do some good here

[00:09:34] but I’m also hearing lots of anecdotes

[00:09:36] to the opposite it’s it’s just not clear

[00:09:39] I’m not sure I can answer your question

[00:09:40] but I I would say just on the principle

[00:09:44] of being a being careful as a business

[00:09:45] person I would probably prefer to er on

[00:09:48] the side of presuming there is like a

[00:09:51] bottom line of value there and there is

[00:09:53] going to be change and I better

[00:09:56] understand this which by the way is also

[00:09:58] the position a consulting firm should

[00:10:00] have because this stuff is so

[00:10:02] complicated and there’s a huge

[00:10:04] likelihood of wasting a bunch of money

[00:10:06] with stuff that goes away in two or

[00:10:08] three years like that’s just the

[00:10:10] scenario you cannot rule out if you’re a

[00:10:11] sensible person that is a huge demand

[00:10:13] for consulting right IT consulties

[00:10:15] should be if not in a position to give

[00:10:17] good answers today, they should be in

[00:10:19] the laboratory working on better answers

[00:10:21] to their question than mine on behalf of

[00:10:23] their clients. There’s a huge need for

[00:10:25] that, I would say, because to your point

[00:10:26] of companies restructuring around AI. I

[00:10:28] see that and I read the headlines and

[00:10:30] honestly it makes me doubt the hard

[00:10:32] value of some of those technologies

[00:10:34] because if I read that certain companies

[00:10:36] are now tying promotions to AI login,

[00:10:39] what that tells me is that the

[00:10:41] technology is not clearly valuable

[00:10:42] enough to be self-promotional. You have

[00:10:44] to force people to use this stuff is

[00:10:46] what the headlines tell us. That was not

[00:10:49] that was not true for the reinvent the

[00:10:52] invention of fire or the wheel or uh uh

[00:10:55] I don’t know cloud computing maybe at

[00:10:58] the very beginning until people have

[00:10:59] reached a certain degree to familiarity

[00:11:01] and then it gets going but to be honest

[00:11:03] chat GPT is a couple of years old

[00:11:05] already generative large language models

[00:11:07] are older expert enough to put a time

[00:11:10] frame on it but some of algorithmic

[00:11:12] stuff was around when I was studying

[00:11:14] linguistics at university which uh god

[00:11:16] forbid is now getting closer to to those

[00:11:18] 15 years. So, it’s been a while. It’s

[00:11:20] been around for a while, right? And if

[00:11:22] if there was genuine superb value in it,

[00:11:24] then that’s that’s my challenge to

[00:11:27] industry, adoption should be way higher

[00:11:28] by now.

[00:11:29] >> Absolutely.

[00:11:30] >> That’s just my feeling.

[00:11:31] >> You made some you made some really

[00:11:32] interesting points there. But so, you

[00:11:34] know, um diving into the other part of

[00:11:36] of this particular topic as a partner

[00:11:39] and the visible authority where your

[00:11:41] entire practice is built around helping

[00:11:43] consultants become recognized and

[00:11:45] trusted experts. What does authority

[00:11:48] actually mean in a world where AI can

[00:11:51] generate insights instantly?

[00:11:53] >> Well, the the word so the word we prefer

[00:11:56] to say is expertise because that’s what

[00:11:57] you get hired for and that is your

[00:11:59] expertise to to stand out as a

[00:12:01] consulting firm. You cannot offer up

[00:12:03] your expertise with zero additional

[00:12:06] constraints or parameters. You need to

[00:12:08] be clear where you apply your expertise

[00:12:11] to solve a specific problem. Honestly,

[00:12:13] that’s everything we do in our business

[00:12:15] is we help consulties make that shift

[00:12:17] going from what we call capability

[00:12:19] selling where you go to market and say,

[00:12:21] “Hey, I have a certain expertise. Who

[00:12:22] here needs that expertise?” Which is the

[00:12:24] pitch of a surprisingly large percentage

[00:12:27] of consulting firms to this day. And

[00:12:29] that is problematic because it puts the

[00:12:31] burden of figuring out how to best use

[00:12:32] your skills on the buyer and buyers are

[00:12:35] not interested in that as much anymore.

[00:12:37] It also uh creates a business problem

[00:12:39] because if you don’t determine the

[00:12:42] constraints of how your expertise gets

[00:12:43] applied that means you’re open to

[00:12:45] everything which means you open your

[00:12:46] business up to a huge degree of

[00:12:48] complexity. Every new client request

[00:12:49] could send you in a different direction

[00:12:51] and make you reinvent the wheel and

[00:12:53] spend more time over the proposal and

[00:12:55] touch up all those slides again. That is

[00:12:57] a complexity cost you do incur if you go

[00:12:59] to market just selling capabilities.

[00:13:00] That’s not my criticism. It’s a fine you

[00:13:02] can do that. Lots of successful firms

[00:13:04] get built through capability selling.

[00:13:06] It’s just there’s a cost to it. That was

[00:13:08] my second point. The previous point was

[00:13:09] the patience and the appetite on the

[00:13:11] buying side to work with you through

[00:13:13] that is coming down and that’s changing.

[00:13:15] So what does authority mean in that word

[00:13:16] or what does it mean or what does

[00:13:18] expertise mean in the word in in in the

[00:13:20] world of AI? We do think that is still

[00:13:22] the only chance you have because if you

[00:13:24] sell capability, oh I I know how to

[00:13:27] write up an operating model redesign.

[00:13:29] Well, chat GPT can do that as well

[00:13:31] because it again it is a form of raw

[00:13:34] capability within a machine and we can

[00:13:36] have a long discussion of how good that

[00:13:38] is, how many errors it makes in

[00:13:39] comparison to your consultant. But the

[00:13:41] two models are very similar indeed. They

[00:13:43] are raw capability as long as they are

[00:13:45] not clearly tied to solving a problem

[00:13:48] and further refined through the

[00:13:49] application of frameworks and experience

[00:13:51] and all that stuff. They are a

[00:13:53] commodity, right? Raw consulting

[00:13:55] capability selling consulting is a

[00:13:56] commodity. Sorry to say it as is a large

[00:13:59] language model. I can pluck that out and

[00:14:01] replace it with a different vendor

[00:14:02] tomorrow. The moment you build authority

[00:14:04] by saying, “Hey, do you know what? I

[00:14:06] don’t just have this capability. I’m

[00:14:07] deploying it to solve a specific

[00:14:09] problem. I give you something specific.

[00:14:10] You can have an SAP consultancy where

[00:14:12] people or or let’s say Salesforce common

[00:14:14] I think where people just say hey we are

[00:14:16] Salesforce consulty. We have we know

[00:14:18] Salesforce. We have 25 certified

[00:14:20] Salesforce experts. Who here needs some

[00:14:22] sales for that?” That’s a commodity. I’m

[00:14:24] sorry to say it. the owners of those

[00:14:26] firms and always say what clients come

[00:14:27] to us because we are flexible and they

[00:14:29] stay with us because they like the team

[00:14:31] and so and I say that’s all true but

[00:14:33] it’s true of every other of the 120

[00:14:35] firms down the street so AI funnily

[00:14:37] enough is the same problem right can do

[00:14:39] everything which means it can do nothing

[00:14:41] which means open AI doesn’t have a

[00:14:42] product as such they have the equivalent

[00:14:44] of electricity but what do you do with

[00:14:46] it so the trick is the flight forward so

[00:14:49] to speak for consulting firms and the

[00:14:50] way how you survive in day of AI is you

[00:14:54] tie your expertise to a specific problem

[00:14:56] and outcome. And this could be just

[00:14:58] making up a joke trying to get unstuck

[00:15:00] from Salesforce, right? We have invested

[00:15:01] in this significantly. It’s gotten a bit

[00:15:04] clunky. We don’t know where the growth

[00:15:05] map is headed. We’d like to get off the

[00:15:07] platform and migrate to something else.

[00:15:09] How do we do that ideally? And the

[00:15:11] moment you say that’s the problem we

[00:15:12] solve in the market and you are the

[00:15:14] first firm to have certified Salesforce

[00:15:16] experts who help clients getting off

[00:15:18] Salesforce, I realize this maybe a bit

[00:15:19] of a weird example, but that’s the the

[00:15:21] point still stands. The moment you

[00:15:23] contextualize your expertise in with a

[00:15:25] certain problem, everything changes

[00:15:26] because now people know what to call you

[00:15:29] for. It’s obvious if I have this

[00:15:30] problem, I need to call this number. You

[00:15:32] get more and more work that’s very

[00:15:34] similar in nature. Every the better you

[00:15:36] are at this, the more the jobs look the

[00:15:39] same, which means now all the variance

[00:15:41] you incur is in the details of the work.

[00:15:44] So now you can pattern recognize and

[00:15:47] dive deeper and find interesting varants

[00:15:51] of the problem you have and do this for

[00:15:53] 3 years now you’re the best in your

[00:15:55] market the region the world of doing

[00:15:57] this specific thing um and in that

[00:15:59] scenario ironically enough not only are

[00:16:01] you better protected against AI because

[00:16:04] that will not have that level of depth

[00:16:06] and its expertise you are even in a

[00:16:08] position where you yourself can make

[00:16:10] better use of it because your expertise

[00:16:13] is now so deep and specific that if you

[00:16:16] prompt the engine, if you built a

[00:16:18] knowledge base of context documents, if

[00:16:21] you fine-tune the algorithm, it’ll be

[00:16:25] incredibly much better and much more fit

[00:16:28] for purpose than someone who doesn’t

[00:16:30] know what you know or just has

[00:16:31] superflous knowledge from having done

[00:16:33] two projects in that space that uses

[00:16:36] chat GPT and is then insanely vulnerable

[00:16:39] uh against hallucinations,

[00:16:41] superfluous answers,

[00:16:44] uh concepts that have been used

[00:16:46] elsewhere, right? It’s funny to me had

[00:16:48] people having chat write their marketing

[00:16:50] plans like by definition that gives you

[00:16:52] something that’s highly probable of

[00:16:54] being used somewhere else. Meaning

[00:16:56] there’s no differentiation baked into

[00:16:57] this at all. So, good luck trying this.

[00:16:59] this idea of building proprietary

[00:17:01] expertise and knowledge being the best

[00:17:04] defensive strategy you can you can play

[00:17:07] and then that even becoming an offensive

[00:17:10] strategy because suddenly you you sit on

[00:17:12] a treasure trove of insight that even

[00:17:14] the large language models don’t have I

[00:17:15] think is the strategy for for this day

[00:17:17] and age

[00:17:18] >> again you made some really really

[00:17:19] interesting points given that there’s so

[00:17:21] much imbalance um you know within

[00:17:24] expertise at the moment across different

[00:17:26] levels how do you then um keep the

[00:17:29] balance with authority, you know, so

[00:17:32] this this road between expertise to

[00:17:35] authority, how how do you keep that

[00:17:37] balance?

[00:17:38] >> Well, I think to to to your point with

[00:17:39] with subjectiveness and flexibility, I

[00:17:41] mean, to a certain point, you’re right.

[00:17:42] It’s always true, right? If something’s

[00:17:44] painful or valuable is always subjective

[00:17:45] to a certain degree, but then I think as

[00:17:48] consultants, uh we are dealing with

[00:17:50] business problems that if I were to ever

[00:17:52] simplify them, all we do as consultants

[00:17:53] is always optimize within the profit

[00:17:55] equation. That’s what it is. Like that’s

[00:17:57] maybe an oversimplification, but that’s

[00:17:58] the idea. And I think there are forms of

[00:18:01] problems, forms of value or outcomes

[00:18:03] people seek and therefore then by

[00:18:05] definition forms of expertise that a are

[00:18:07] almost objectively real and true. And B,

[00:18:10] if you’re good about what you do, they

[00:18:12] might be persistent. Not forever, but

[00:18:15] sticking with my silly example of

[00:18:16] getting off of Salesforce. I mean, by

[00:18:18] definition, that’s a market that if you

[00:18:20] the more successful you are, the sooner

[00:18:22] it comes to its end. But I don’t think

[00:18:24] that you will see the end of it in your

[00:18:25] 20 person firm over the next 3 or 5

[00:18:27] years. So I do think that is that is

[00:18:29] pretty clear, pretty clearcut and and

[00:18:31] pretty persistent in that sense. And um

[00:18:34] to I think if I understand your question

[00:18:35] correctly, how do I become how do I go

[00:18:38] from being an expert in that field to

[00:18:41] being an authority? Well, you talk about

[00:18:43] your learnings a lot and you educate

[00:18:46] people about why the problem you solve

[00:18:48] is relevant. Meaning you need to

[00:18:51] contextualize what does it do to a

[00:18:52] business if it doesn’t solve that. You

[00:18:54] could have perspectives on how this

[00:18:56] should be addressed and maybe not. You

[00:18:58] should be able to help people make the

[00:19:00] case for a change. Right? This is the

[00:19:02] question of why should we address I I

[00:19:04] understand the problem. I see what it

[00:19:05] does to my business but why would I

[00:19:06] change now? What’s the argument I have

[00:19:08] to give to my boss? Like all these

[00:19:09] things you should be completely

[00:19:11] outspoken and public about that. I often

[00:19:14] say the stance and the process a

[00:19:16] consultancy needs to take all it needs

[00:19:18] to do in terms of its marketing is the

[00:19:19] stance of academic publishing. Once

[00:19:21] you’re clear on what the problem is you

[00:19:23] you focus on which is the same

[00:19:24] researchers do. You need to go in a lab

[00:19:27] and figure this thing out while at the

[00:19:29] same time constantly publishing about

[00:19:31] it. That’s it. It’s almost a mirroring

[00:19:34] of the scientific process. And I think

[00:19:36] if then your observation on the market,

[00:19:39] your entrepreneurial bet is true, the

[00:19:41] bet, that bet being that this problem is

[00:19:43] not only real, but it’s experienced as

[00:19:45] painful and urgent for a large enough

[00:19:47] group of companies to give to make a

[00:19:49] market for you. If that is true and you

[00:19:51] publish consistently in that space with

[00:19:54] hopefully ever more interesting

[00:19:55] insights, it’s you almost can’t help but

[00:19:59] but becoming an authority in the place.

[00:20:01] I mean, sooner or later there will be

[00:20:03] competition. There will hope there

[00:20:05] usually there is already some otherwise

[00:20:08] the market is not real and then if you

[00:20:10] are successful people will try and

[00:20:11] follow in your wake but usually that

[00:20:14] that room still will be big enough for

[00:20:16] you to still become an authority in your

[00:20:18] own right maybe not the biggest maybe

[00:20:21] not the worldwide version but in your

[00:20:23] regional market in your sphere of

[00:20:24] influence in your specific variant of a

[00:20:28] bigger problem like you can still be the

[00:20:29] authority and I think that is the job is

[00:20:31] to push for that going back to earlier

[00:20:33] questions in the conversation I think AI

[00:20:35] increases the pressure of doing that.

[00:20:38] That pressure was always there. You were

[00:20:40] always better off being a consulting

[00:20:42] firm with a clear proposition and

[00:20:43] expertise as opposed to capability

[00:20:46] selling. But it was very viable to build

[00:20:48] a company grow 7% a year with capability

[00:20:50] selling. I think the increasing

[00:20:53] automation and softwareization of

[00:20:55] knowledge work, it’s not just AI, right?

[00:20:57] Let’s be honest. It’s also SAS solutions

[00:20:59] like like an ERP that also generates a

[00:21:02] dashboard and a cash flow projection

[00:21:04] that was work that used to be

[00:21:05] consulting. I mean go back further in

[00:21:07] time people would employ 20 juniors to

[00:21:09] draw up a spreadsheet before spreadsheet

[00:21:12] software came along like you have

[00:21:13] software eating at those types of

[00:21:15] forever. So that pressure to get to push

[00:21:17] ever more into this expertiseled don’t

[00:21:20] just sell my business partner Luke

[00:21:22] smiles and I we always say don’t just

[00:21:24] sell me the means. So don’t don’t sell

[00:21:26] me I know Salesforce. Sell me the end. I

[00:21:28] want to get off Salesforce. I want to

[00:21:30] rein in my spend in Salesforce. We want

[00:21:32] to scale up our sales for whatever it

[00:21:34] is, right? But I’m interested in the end

[00:21:36] and not the means. And that pressure to

[00:21:38] those things exist on a scale and moving

[00:21:40] from the means towards the end and then

[00:21:42] the end after that and then the end

[00:21:43] after that that is I think extremely

[00:21:45] intensified through the advent of AI and

[00:21:48] and the hype around AI both. Also Floren

[00:21:51] on your LinkedIn you make this point

[00:21:52] that all consultancies sound generic

[00:21:55] when they describe themselves but become

[00:21:57] distinctive when they describe the

[00:22:00] issues that they solve which I think

[00:22:02] most of the consultancies are trying to

[00:22:03] get into or they try to get there. Now

[00:22:06] if you know AI can generate polished

[00:22:09] propositions for any firms in 30 seconds

[00:22:11] which most of the clients also do now

[00:22:13] before hopping into the calls you know

[00:22:15] they are more prepared these days. So

[00:22:17] doesn’t that mean the only consultancies

[00:22:19] that survive are the ones that have gone

[00:22:22] so deep into one specific client problem

[00:22:26] that no AI can replicate you know what

[00:22:29] they know which I think again is very

[00:22:32] difficult in this space.

[00:22:34] >> So yeah again nuanced answers to this.

[00:22:37] The first thing is I think that what I

[00:22:40] what I meant when I said everybody

[00:22:41] sounds the same because they’re all

[00:22:42] offering their capabilities like what

[00:22:44] what is a Salesforce consultancy to say

[00:22:46] other than we know Salesforce or

[00:22:48] versions thereof until they shift and

[00:22:50] they go issue and proposition first. Now

[00:22:53] they’re already different from all the

[00:22:54] other capability sales but step one but

[00:22:56] this is of course just the beginning and

[00:22:57] I think this is where you’re right even

[00:22:59] if you were to say we solve this problem

[00:23:01] there could be 20 other firms saying the

[00:23:03] same tomorrow. What I think my post

[00:23:05] you’re referring to said was once you

[00:23:07] start with the issue in mind now you

[00:23:10] have the chance to get really different

[00:23:11] from others because you can ask all

[00:23:13] kinds of interesting questions. Sticking

[00:23:15] with my silly example of getting off of

[00:23:17] Salesforce. You can start to ask when

[00:23:19] does that even make sense for me as a

[00:23:21] client? How would you the consultancy

[00:23:23] determine that it’s a smart decision in

[00:23:25] my case? What are the best steps to take

[00:23:27] first? Among those best steps, how do

[00:23:30] you decide which one? Like do you know

[00:23:31] what I mean? Starting from the issue,

[00:23:33] you can inquire your own way of working

[00:23:35] with firms, with clients to find those

[00:23:39] bits and pieces of differentiation and

[00:23:41] then you can stack them on top of each

[00:23:43] other. So we always talk about how

[00:23:45] differentiation is built brick by brick

[00:23:46] of all the problems in the Salesforce

[00:23:48] universe you could solve. You pick one

[00:23:50] or if you’re a larger firm, a category

[00:23:53] of one and then you have subpropositions

[00:23:56] as we call them. So more specific

[00:23:57] problems underneath that. So you pick

[00:23:59] that. Okay, now you’re already different

[00:24:00] from people who didn’t make that choice.

[00:24:02] Then you say, “Here’s our logic for

[00:24:04] determining whether that’s a helpful

[00:24:06] step or not.” You could argue with that,

[00:24:07] but you now you’re clear. You have made

[00:24:09] a choice and that that means you’re

[00:24:10] different. So, it’s the stacking of

[00:24:12] choices that take you away from

[00:24:13] competition. And as long as the

[00:24:16] throughline and the business rationale,

[00:24:19] the degree to which those choices make

[00:24:21] sense are there and they’re visible and

[00:24:23] they’re real and you can have clients

[00:24:25] that say, “Yeah, I buy that.” you always

[00:24:27] have an argument to stand on and I think

[00:24:30] that is how that works in principle. Now

[00:24:32] to your point of yeah but AI could do

[00:24:35] that. I’m not discussing that maybe

[00:24:37] sometimes it could. I’m 100% positive it

[00:24:40] cannot do that today because it is a

[00:24:45] probabilistic token system right it

[00:24:47] comes up with highly probable answers

[00:24:50] which are not where specific specificity

[00:24:53] and distinction and actually also

[00:24:55] expertise are found because very often

[00:24:57] if you think about it real expertise

[00:24:59] often comes down to knowing when not to

[00:25:01] apply the established best practice

[00:25:03] experience has to do a lot with

[00:25:05] exceptions to rules that’s a breaking

[00:25:07] point for AI already. So the second

[00:25:09] question then is even if AI can do a lot

[00:25:12] of it and there might be cases where it

[00:25:14] can. Again I’m not a software engineer.

[00:25:15] The way that technology works today,

[00:25:17] you’ll still need a person in the mix to

[00:25:20] supervise the AI to take legal liability

[00:25:24] to fix something if that goes wrong. Do

[00:25:27] you want to do that? Do you want to have

[00:25:29] that person on staff all the time or

[00:25:31] would you rather hire an external agency

[00:25:34] that has those expert continuously

[00:25:36] trains them and involves their knowledge

[00:25:37] as the tools changes which that’s the

[00:25:39] definition of a consulting firm I think

[00:25:41] you’d rather go with a consultancy so

[00:25:43] I’m not too concerned about the

[00:25:45] technology if you have a strong problem

[00:25:48] solution orientation so if you are

[00:25:50] propositionled consultancy as we call

[00:25:52] this and a lot of the questions we get

[00:25:54] around that stem from the history of

[00:25:57] consulting to build by the hour if I can

[00:25:59] shine a light on that because software

[00:26:02] as a service never has that question

[00:26:04] they just have the confidence to never

[00:26:05] discuss their pricing right we talk

[00:26:07] about all this oh if the computer does

[00:26:08] the work if it goes away whatn not whatn

[00:26:10] not well if you make the mental shift of

[00:26:12] saying I’m actually solving this problem

[00:26:15] and you pay me by the sol you pay me for

[00:26:17] having the problem solved the hours

[00:26:19] don’t factor into this whether my AI did

[00:26:21] it or not fac doesn’t factor into this

[00:26:24] whether your AI could do the same in

[00:26:26] theory doesn’t factor into this. I mean,

[00:26:28] low code tools are a thing. I could

[00:26:30] build my own website. No problem. Of

[00:26:32] course, we still have a service provider

[00:26:34] because why would I waste my hours

[00:26:36] learning that why that doesn’t the

[00:26:38] business case is not there. So, the

[00:26:39] moment you step away from this billing

[00:26:41] by the hour logic and you you look at

[00:26:42] for example software as a service, if I

[00:26:44] can call out Microsoft for a minute,

[00:26:46] they are charging me the same amount for

[00:26:48] PowerPoint and the maintenance of the

[00:26:50] PowerPoint service forever. They make a

[00:26:52] very meaty profit margin on it. They

[00:26:55] haven’t improved the damn software for a

[00:26:57] decade. Like I have the same issues with

[00:26:58] PowerPoint that I use it a lot. I’ve had

[00:27:00] 10 years ago. So there’s no improvement

[00:27:02] in quality. And yet, not only do they

[00:27:04] charge me every month, the prices go up

[00:27:06] all the time. I’m not saying that’s

[00:27:08] great business practice. In fact, I

[00:27:09] think the opposite is true. All I’m

[00:27:10] saying is consultants could maybe borrow

[00:27:12] a page from that book and come back to,

[00:27:14] and I said it before, I solved this

[00:27:16] problem. We have built all kinds of

[00:27:18] methodology and process to increase the

[00:27:20] actual chance of this problem being

[00:27:22] fully solved and you getting towards

[00:27:23] that outcome. I mean we don’t control

[00:27:24] the outcome in full as consultants

[00:27:26] never. But we can maximize the chances

[00:27:28] of a client’s success. What’s that worth

[00:27:30] to you? And there’s no did an AI do it,

[00:27:33] did the intern do it. That discussion is

[00:27:35] that is not part of that discussion. It

[00:27:37] is the question of do I as the client

[00:27:38] get the problem solved and do I have

[00:27:40] high confidence in your ability of doing

[00:27:41] it? And that I get that that’s a

[00:27:43] fundamental shift of not just the

[00:27:45] mindset but then also some internal

[00:27:46] structures within consulting. But those

[00:27:48] are manageable that sounds are going to

[00:27:50] know. But we have clients that did it.

[00:27:52] So I know it’s possible and I think that

[00:27:55] is the work so many firms have in front

[00:27:57] of them right now right because of the

[00:27:59] changes that I’m not going I’m not

[00:28:00] denying that there’s a reset in the

[00:28:02] market because of AI. That is really

[00:28:04] true for many parts

[00:28:05] >> yeah of the sector. So, so Florian, in

[00:28:07] in retrospect, you know, if clients now

[00:28:10] rely on consultants less for answers and

[00:28:13] more for validation and decision and

[00:28:16] confidence, do you think that it means

[00:28:18] that AI actually elevates the best

[00:28:21] consultants?

[00:28:22] >> I think we see that happening already.

[00:28:23] There seems to be a benefit that goes to

[00:28:26] smaller, very, very narrowly positioned

[00:28:28] expertise boutiques that have more

[00:28:30] senior people because they have the

[00:28:32] expertise in spades. They always had it.

[00:28:34] They never could get into the can you

[00:28:37] send 82 developers to 20 of my sites

[00:28:40] tomorrow business. That was always with

[00:28:41] the big four with the highly leveraged

[00:28:43] firms. They are having more

[00:28:45] opportunities to compete now because

[00:28:46] they can say no I cannot send you 80

[00:28:48] people but I have one very senior

[00:28:50] business solution architect. He comes

[00:28:52] equipped with 20 agents and a sort of

[00:28:54] process we’ve built in software here. He

[00:28:56] can do the same amount of work. So what

[00:28:58] I said before that capacity can reside

[00:29:00] in a computer seems to put the big guys

[00:29:02] who built all the leverage in human

[00:29:04] brains under pressure while it seems to

[00:29:07] benefit the smaller guys who always had

[00:29:09] the brains but not the the or who was

[00:29:11] had the expertise but not the leverage

[00:29:12] in sort of numbers of brains and pair of

[00:29:14] hands and whatever it is. So that that

[00:29:15] seems to be real and I do think concern

[00:29:18] AI can be an amplifier and if going back

[00:29:21] to what I said earlier if Javon’s

[00:29:23] paradox is real what we might see is the

[00:29:26] relevance and the cost and all that

[00:29:28] stuff in the value chain like how many

[00:29:30] people do I need to generate a

[00:29:32] consulting outcome that could come down

[00:29:33] right with that prices come down to a

[00:29:36] certain extent despite what I’ve said

[00:29:37] before that you could price on the

[00:29:39] what’s the outcome worth to a client but

[00:29:41] prices could on average because maybe

[00:29:43] just maybe someone who says Oh, I also

[00:29:45] need this outcome. It’s worth only

[00:29:47] $20,000 to me. I don’t have more because

[00:29:49] I’m a small mom and pop shop and I would

[00:29:51] never have bought consultancy in a

[00:29:53] previous life. But it could now be worth

[00:29:54] the consulty’s time because they could

[00:29:56] say, “Yeah, we could send you a midlevel

[00:29:58] consultant and two agents and he could

[00:30:00] do that and we could make a nice profit

[00:30:01] on those 20 grand.” It would have been

[00:30:03] cost prohibitive to work for you guys in

[00:30:05] the future. But now it’s possible like

[00:30:07] that is also an opportunity that opens

[00:30:08] up and I don’t know where that goes to.

[00:30:10] Does that go to the small players

[00:30:11] exclusively because they are closer to

[00:30:13] that small midsize market or do the big

[00:30:15] guys push for that? I don’t know. It

[00:30:17] might level out. But between the the

[00:30:20] unit economics changing, let’s put it

[00:30:22] that way, right? And then that impacting

[00:30:25] demand, it could level out in all kinds

[00:30:27] of interesting ways. But but AI will

[00:30:29] definitely amplify fasttrack at capacity

[00:30:32] where that didn’t exist. And honestly,

[00:30:34] maybe even dare I say improve quality.

[00:30:37] Um I don’t trust any of the models

[00:30:39] further than I can throw them as we have

[00:30:41] a weird idium that works like that in

[00:30:43] German which is a way of saying I don’t

[00:30:45] trust them much but there are ways in

[00:30:47] which they work well and improve quality

[00:30:50] of a consulting outcome in that I can

[00:30:52] now do things I would not have done

[00:30:55] before because AI completely collapsed

[00:30:57] the timeline to give one specific

[00:30:59] example how we are using it. We used to

[00:31:02] mock up a website for workshops as as

[00:31:06] the proposition work comes together for

[00:31:08] clients because it’s a very tangible way

[00:31:10] for people to feel what the story and

[00:31:12] the service architecture might look like

[00:31:14] and we used to mock them up in

[00:31:16] presentation software which is a manual

[00:31:17] process and that meant we were we would

[00:31:19] give people a new homepage so they can

[00:31:21] get the idea. Nowadays we go into a

[00:31:23] noode website generation tool and we

[00:31:25] build them an entire website. They can

[00:31:27] walk through this they can criticize it.

[00:31:29] we can make changes on the thing which

[00:31:30] gives them a much better feeling not

[00:31:32] just how it would would work but it

[00:31:33] enabled us to have workshops about where

[00:31:35] could this break we have redesigned the

[00:31:37] proposition you still have this old

[00:31:39] service over here we not sure it fits

[00:31:41] like that becomes way more we could

[00:31:42] never have done that like we could if we

[00:31:45] had priced it before clients would not

[00:31:47] have paid for it if we had done it at

[00:31:48] our own cost the margin would have

[00:31:50] evaporate now it really improves the

[00:31:51] quality of the product because at a very

[00:31:54] low low marginal cost I can add that

[00:31:56] onto the the service I’m already I I

[00:31:59] expect to see a lot of that also.

[00:32:01] >> Yeah. And so Florian on the on the data

[00:32:03] privacy um angle, you know, um AI

[00:32:06] powered consultancies require clients to

[00:32:08] hand over sensitive data to um firms uh

[00:32:12] using different tools which they don’t

[00:32:13] even fully understand at the moment. So

[00:32:16] is that creating a new trust barrier

[00:32:19] that probably no one is talking about?

[00:32:21] >> Well, I can’t speak for everybody’s

[00:32:22] clients. We certainly sign if if you get

[00:32:26] into the Fortune 200, Fortune 100 size

[00:32:28] clients, they have an AI governance.

[00:32:31] They will make you sign waiverss and

[00:32:33] stuff and they will check on that and if

[00:32:35] they catch you putting a question that

[00:32:37] sort of has traces of their name and

[00:32:39] chat GPT, you know where that goes. So

[00:32:41] yes, there’s absolutely not only is this

[00:32:44] creating a an issue of trust, corporate

[00:32:46] clients are already working on the

[00:32:48] solutions and they are thorough with

[00:32:50] that. So that that is an additional cost

[00:32:52] an additional burden that’s coming to

[00:32:54] smaller consultancies in particular

[00:32:55] because you’ll the job of mitigating

[00:32:58] those concerns and complying with those

[00:33:00] governance policies that that’s on us.

[00:33:02] There are ways of working around this

[00:33:04] and AI can play a role right game we

[00:33:06] play here is have a locally running

[00:33:08] model to sanitize inputs and then if you

[00:33:10] need to consume a public cloud service

[00:33:12] through an API make it’s the first step

[00:33:14] sanitize the inputs then run it in the

[00:33:16] cloud and then play that backwards. they

[00:33:18] can be technologically a bit more um

[00:33:20] difficult to set up which then again

[00:33:22] gives you an edge as the consulting firm

[00:33:24] because you have that setup and you

[00:33:25] could be the instance that the client

[00:33:27] goes through to have some of those

[00:33:29] questions answered with the help of a

[00:33:31] machine. They themselves might not be

[00:33:32] willing to take the same steps while

[00:33:34] also being prevented from feeding their

[00:33:36] entire production data into or copies

[00:33:39] thereof into into any into into any

[00:33:42] because I mean two consultancies have

[00:33:44] been in the press for being hacked with

[00:33:45] this stuff already, right? You cannot

[00:33:46] like this. You absolutely have to have

[00:33:49] good safety procedures there at all

[00:33:52] levels. And then you got to doublech

[00:33:54] check the the outputs once they come

[00:33:55] back to you. Right. It’s is really and

[00:33:57] these are this is also why I have

[00:33:59] questions about the limits of those

[00:34:00] technologies because there are these

[00:34:02] constraints and additional costs already

[00:34:04] and they make sense from a commercial

[00:34:05] perspective. Now it’s still very

[00:34:07] attractive to use the models because the

[00:34:09] token costs are subsidized by venture

[00:34:11] capital funding. That’s where we are

[00:34:12] today. What happens if that funding goes

[00:34:15] away tomorrow? Will you still embed

[00:34:17] cloud code at at the core of your

[00:34:19] business processes? I I’m wondering I

[00:34:22] don’t know but I’m wondering which is

[00:34:23] why have why have why I have real

[00:34:25] questions about the seriousness of IT

[00:34:27] consultancies that just without any

[00:34:28] questions asked go in and say hey we are

[00:34:31] AIE AI first this AI first that this to

[00:34:34] me um as a buyer and I am at a very

[00:34:36] small scale sometimes a buyer of these

[00:34:38] things is a severe red flag. a

[00:34:40] consulting firm that comes in and says,

[00:34:41] “Hey, AI opens a few possibilities or

[00:34:44] opportunities in your field which we

[00:34:46] have seen work great and these are the

[00:34:48] constraints within which you could seize

[00:34:49] them and here’s how we make that.”

[00:34:51] That’s a much better pitch. Like again

[00:34:53] to your earlier questions, expertise and

[00:34:55] experience and the value of that is it’s

[00:34:57] all about nuance and differentiation and

[00:34:59] these overgeneralizing flat statements.

[00:35:01] Oh, AI this, AI that. No, we see where

[00:35:03] that ended. It ends with Open Cloud

[00:35:05] exposing all your API keys to the world.

[00:35:07] I’d advise any consultancy against those

[00:35:10] practice. I I get the appeal. You can

[00:35:11] make lots of money right now because

[00:35:13] there’s hype in the market and you can

[00:35:14] capture that demand. Can you build a

[00:35:16] sustainable practice on the back of

[00:35:18] that? I don’t know. But I have a

[00:35:20] question. Anyone remember the metaverse

[00:35:23] consultants and the practices of the big

[00:35:24] houses and the trillions of dollars in

[00:35:26] metavverse revenues that very reputable

[00:35:28] consulties predicted would exist. What

[00:35:30] happened to those guys? I can tell you

[00:35:32] that many of them are not working for

[00:35:33] those firms anymore because these

[00:35:34] practices got dissolved. So don’t make

[00:35:36] the same mistell you many of them by

[00:35:38] buying too much of the hype. Take a

[00:35:41] constructively critical stance and

[00:35:43] advise people on how that technology

[00:35:45] makes sense for them. That’s a much

[00:35:46] better ticket I would say.

[00:35:48] >> Yeah. Yeah. And now if we dive into who

[00:35:50] survives and how you know in in in this

[00:35:52] time and age if the traditional billing

[00:35:55] model is under pressure which I think it

[00:35:57] already is what new models do you see

[00:36:00] emerging? Is it either subscription

[00:36:02] outcomebased or AI enabled platforms?

[00:36:05] >> So if you can build a platform if you if

[00:36:07] you build this mixed model which I don’t

[00:36:09] know enough about where Palunteer seems

[00:36:12] to be sort of a bit of a trade. So

[00:36:14] combining proprietary software

[00:36:16] technology with a service that’s

[00:36:18] embedded at the client. I think that’s

[00:36:20] when you can get subscription-like

[00:36:22] recurring revenue business models. I I

[00:36:24] don’t understand those very well because

[00:36:25] I’m not deep enough in the software

[00:36:26] world but they are proven already. I

[00:36:28] mean take look at Ponty, right? So I

[00:36:29] think that is one direction you can go

[00:36:31] for many smaller firms. The same is true

[00:36:33] if you’re running managed services as a

[00:36:34] as an IT consulty that that certainly

[00:36:37] makes sense. And I think you can again

[00:36:39] smaller players maybe can now enter

[00:36:41] plays that were cost prohibitive for

[00:36:43] them before like if I buy an outsourcing

[00:36:45] deal with a large provider take take a

[00:36:47] TCS in India or take a an Accenture or

[00:36:50] those firms if they if I buy a managed

[00:36:52] service from them I’ll often get a cost

[00:36:54] decrease guarantee in my contract. So

[00:36:57] over 3 years they’ll they’ll lower cost

[00:36:59] per ticket 5% every year. Making those

[00:37:01] numbers up, right? Because obviously

[00:37:02] they have a delivery capacity center

[00:37:05] somewhere in in a lower labor cost

[00:37:07] geography. They have rigorous processes,

[00:37:09] automation, all that stuff to drive

[00:37:11] their own delivery costs down. Could I

[00:37:13] mimic that as an onshore only 30 person

[00:37:16] consultancy in mid Europe in the past?

[00:37:18] Not necessarily. Maybe now with AI I

[00:37:20] can, right? Sorry that that’s a

[00:37:21] challenge. But so so maybe even in those

[00:37:23] proven models some of the commercial

[00:37:25] details might change. Uh moving further

[00:37:27] down the chain the big thing from my

[00:37:30] perspective right now and we see that

[00:37:31] with a lot of our clients is fixedric

[00:37:34] programs which that’s the first thing

[00:37:35] that opens up as you are a proposition

[00:37:38] consultancy that specializes in solving

[00:37:40] a problem. You still sell time and

[00:37:42] materials for a little bit until you

[00:37:44] have pattern recognition. You say, “Hey,

[00:37:46] every time to get back to your example,

[00:37:48] every time a dentist comes in with this

[00:37:49] particular problem for a website, we do

[00:37:51] those five steps and we take on average

[00:37:53] 14 hours. So, let’s fix price this thing

[00:37:55] at 16 hours, our internal cost rate and

[00:37:57] then see if we can drive efficiencies to

[00:37:59] improve the margin on that.” That’s I

[00:38:02] think the best ticket. It’s also very

[00:38:03] easy for clients to buy because they

[00:38:05] understand this is the structure. I get

[00:38:07] this. These are the deliverables. This

[00:38:08] is the fixed price. Very nice. And if

[00:38:10] you do enough projects that are very

[00:38:12] similar, which you should be doing if

[00:38:13] you are propositionled, you can

[00:38:15] statistically sort of reduce the risk of

[00:38:17] that over time, pretty quickly usually

[00:38:18] even. Um, so that’s a good way of going.

[00:38:20] Outcomebased or valuebased pricing, uh,

[00:38:23] I read a lot about that in the press.

[00:38:25] McKenzie does it and so forth. It’s been

[00:38:27] around ever since I’ve been in

[00:38:28] consulting and it’s never been the

[00:38:30] default model and I have doubts about it

[00:38:33] becoming so. Why? It’s just structural

[00:38:35] reasons. procurement doesn’t like it

[00:38:37] because what’s the outcome based pricing

[00:38:39] on a website? How much of that do you as

[00:38:42] a consultancy control versus the client

[00:38:43] controls it? How are we going to measure

[00:38:45] that? What type of contractual agreement

[00:38:48] and data sharing needs to be in place

[00:38:50] for you, the consultancy, to be able to

[00:38:52] check my homework? The visitor numbers

[00:38:54] are mine. That’s my data. You’re not

[00:38:55] seeing it. So, if I’m telling you, oh,

[00:38:57] there’s no been no increase in website

[00:38:59] visits. How are you going to invalidate

[00:39:01] that to to drag me in front to to

[00:39:03] enforce the outcome? Right? So it

[00:39:05] creates all kinds of complications that

[00:39:07] the moment someone comes in and say,

[00:39:08] “Hey, how about we say $120 an hour?”

[00:39:10] Procurement is going to jump on those

[00:39:12] guys because they don’t want to deal

[00:39:13] with the headaches. And that’s just the

[00:39:15] the the legal or compliance side of it.

[00:39:17] There are technical challenges as well.

[00:39:19] Many consultancies drive improvements

[00:39:21] that a client is in no position to even

[00:39:24] do a baseline measurement on as the work

[00:39:25] begins. Right? Many consultancies have

[00:39:27] to clean up the space to measure a

[00:39:29] baseline at the outset. So can the

[00:39:31] client even ever measure the outcome

[00:39:33] they’re getting? Are they in that

[00:39:35] position? Yes or no? So there’s all

[00:39:36] kinds of complications and that’s that’s

[00:39:39] why I have doubts about it. That’s also

[00:39:40] the reason why if you look in those

[00:39:43] large consultancy does outcomebased

[00:39:45] deal. Yeah. When I was still at

[00:39:47] Accenture, we did a couple of them. We

[00:39:48] usually did them in the form of joint

[00:39:50] ventures, right? These are not your

[00:39:52] run-of-the-mill typical consulting

[00:39:53] agreements. These are contracts that set

[00:39:56] up a new business. You own 50%, I own

[00:39:59] 50%, both of us look into all the

[00:40:01] numbers. It’s very clear-cut in terms of

[00:40:03] profit splitting and that that is how

[00:40:04] you can run those things and it can

[00:40:06] really be outcome or value based, but

[00:40:08] that doesn’t lend itself to every type

[00:40:10] of consulting. So yeah, I read the other

[00:40:13] day Mckenzino does 25% of a based

[00:40:15] prices. Yeah, they do. I believe the

[00:40:18] headline, but maybe the structure, the

[00:40:20] logic and the frameworks they can apply

[00:40:22] might be cost prohibitive for smaller

[00:40:23] shops. So I I don’t know about the the

[00:40:25] the value and the outcome based. What I

[00:40:28] do think might be there is a sort of a a

[00:40:31] bit of a more that is more similar to

[00:40:33] risk sharing, right? Where you do a mix

[00:40:35] of hey, it’s a fixed price engagement

[00:40:37] and there’s a bit of a kicker if we see

[00:40:38] certain metrics or there’s a setup fee

[00:40:41] and then we collect this fixed sum if we

[00:40:45] see certain that there there are

[00:40:46] varieties of this where you take risk

[00:40:48] off the client side. I am personally as

[00:40:51] a partner in a business not always a fan

[00:40:54] of those because if there is a power

[00:40:56] imbalance in the relationship as in the

[00:40:58] consultant is much smaller than the

[00:41:00] client for example and you have

[00:41:02] corporate politics applying on the buyer

[00:41:04] side these risk models might create

[00:41:07] weird incentives because if I’m not at

[00:41:10] the sponsor or highest buyer level but

[00:41:12] I’m the department head I might have an

[00:41:14] incentive to sabotage your work if your

[00:41:16] profit kicker comes out of my budget.

[00:41:18] Does that make sense? So there can be

[00:41:20] I’m not saying clients are doing that.

[00:41:22] I’m not saying it happens a lot. I’m

[00:41:23] just saying again there are

[00:41:24] complications to those models that might

[00:41:27] make it so that you at the end of the

[00:41:28] day said you know what fixed price is

[00:41:29] much better to do and to deal with on

[00:41:31] both sides. So let’s go with that.

[00:41:34] That’s just my personal experience.

[00:41:35] >> You’ve spent nearly 5 years inside

[00:41:38] right? So which is one of the most

[00:41:40] advanced tech consultancies in the

[00:41:42] world. From that experience, Florian,

[00:41:43] what do you think that big firms like

[00:41:45] Assenture are doing right that smaller

[00:41:48] consultancies are simply not doing or

[00:41:51] vice versa?

[00:41:52] >> Well, oh, good question. So, let me

[00:41:54] start by saying these are fundamentally

[00:41:55] different games. Take the fact that

[00:41:57] within my first couple of weeks at

[00:41:59] Accenture as in a global marketing role,

[00:42:01] I had to learn that marketing’s job is

[00:42:02] not to win net new clients in the

[00:42:04] business I was in. That was not the job

[00:42:06] because that part of the business wanted

[00:42:07] to work for the Fortune 500 and had at

[00:42:09] the moment in time when I joined 500 of

[00:42:11] those clients. So right that’s not a

[00:42:12] situation as a midsize consultant like

[00:42:14] net lo net new logo acquisition not

[00:42:16] exactly a job. So there there’s big

[00:42:19] differences between those guys and the

[00:42:20] size they have also in terms of the

[00:42:22] business they are in they seem like they

[00:42:24] are broad and do everything for

[00:42:25] everybody whereas in reality and this is

[00:42:28] maybe something you can copy. They are

[00:42:30] specialized in a weird way. They are

[00:42:32] specialized in the leveraged capacity

[00:42:33] model which is a type of consulting work

[00:42:35] your 30 person firm doesn’t get to do.

[00:42:37] Yes, technically you might be doing the

[00:42:39] same work and yes, you might even be

[00:42:40] smarter in Salesforce or your expertise

[00:42:42] might be deeper. But if a client says, I

[00:42:45] have 162 sites, I need teams of 20 on

[00:42:48] half of them. Can you start next week?

[00:42:50] You are already out of the picture and

[00:42:52] it’s only the big four and a couple of

[00:42:53] technology that is their market that is

[00:42:55] their business and you best presume that

[00:42:58] it’s very different from yours although

[00:43:00] there are very similarities

[00:43:02] also but that sending that ahead. So

[00:43:04] that said, what the big guys are good at

[00:43:06] obviously is everything that has to do

[00:43:09] with standardizing, scaling, managing,

[00:43:13] running. So operational excellence is a

[00:43:16] huge one. Um, not just in delivery,

[00:43:20] streamlining that, but also in the

[00:43:22] things that sometimes the boutique

[00:43:23] owners tend to find a bit more boring.

[00:43:25] Uh, financial management of the ins and

[00:43:27] outs, all that stuff. Obviously, they’re

[00:43:29] they’re very very sophisticated in those

[00:43:31] spaces, but that’s kind of obvious. I

[00:43:33] think a less obvious one is the degree

[00:43:35] of sophistication in propositionled

[00:43:37] selling that these firms do and you

[00:43:39] could never tell because the marketing

[00:43:41] you see from the outside the posters on

[00:43:43] the airports and the websites and stuff

[00:43:45] that is generic capability I’ll say it

[00:43:48] my business partner Luke will say the

[00:43:50] same fluff you should not copy at all

[00:43:52] but that is a function of their size

[00:43:54] they are so big it’s very difficult to

[00:43:56] have any specificity at that higher

[00:43:58] level but you better not believe that

[00:44:00] they don’t have that specificity

[00:44:03] They have very very clear very very

[00:44:06] tailored proposals. We used to call them

[00:44:09] sellable units in Accenture. So packages

[00:44:11] of service offerings and people and so

[00:44:13] and these things come with propositions

[00:44:16] and messaging and architecture and

[00:44:18] delivery logic that is very very

[00:44:20] refined. You might not see it because

[00:44:22] you’re not the buyer but they exist. And

[00:44:24] I don’t know thinking back of the

[00:44:26] examples in I was in an in extent

[00:44:28] industry X business they sell digital

[00:44:30] transformation to industrials. They had

[00:44:32] very very specific offerings for site

[00:44:35] turnaround for mid-stream refineries in

[00:44:37] the US. I’m not going to tell you what

[00:44:38] was in those packages, but we’re talking

[00:44:40] at that level and they were very clear

[00:44:41] about these are the five biggest

[00:44:43] problems for pipelines in that part of

[00:44:46] the ocean or what like to it gets to

[00:44:48] that level of specificity and um I get

[00:44:51] that is a bit of pointless advice

[00:44:52] because I’m telling you copy the stuff

[00:44:54] you can’t see but but I think just just

[00:44:57] take maybe take my word for it that

[00:44:59] logic is there. They’re very big on the

[00:45:01] capability selling in those closer to

[00:45:04] the revenue parts of their business,

[00:45:05] those parts of business development. And

[00:45:07] as a mid-size consultancy, you should

[00:45:09] copy this and this exclusively, not the

[00:45:12] brand marketing they’re doing. That is

[00:45:14] for later. You save that for for when

[00:45:16] you have become successful and big and

[00:45:18] you have the margins. Do the other

[00:45:20] stuff, the stuff that’s much closer to

[00:45:22] the talking to a specific client. Now to

[00:45:25] addressing those pain points being then

[00:45:27] building up from there until you get

[00:45:28] more and more abstract and general about

[00:45:31] these things that’s just that’s a

[00:45:32] function of their size. But the way you

[00:45:34] win in the market is you have a very

[00:45:36] specific proposition. You’re very clear

[00:45:38] on who you put that in front of and you

[00:45:41] build the internal piping to deliver

[00:45:42] that very smoothly. Those are the

[00:45:44] ingredients.

[00:45:45] >> And what do you think which firms are

[00:45:47] most at risk at the moment? Is it the

[00:45:50] large incumbents um midsize

[00:45:52] consultancies or niche players?

[00:45:54] >> Uh the I think the middle is is will get

[00:45:57] the brunt of it because the the large

[00:45:58] guys they’re on the press a lot. They

[00:46:00] are laying off people. They’re clearly

[00:46:02] under pressure. They’re also investing

[00:46:04] capital to ramp up AI capabilities.

[00:46:05] They’re buying shops. So they have deep

[00:46:07] pockets and they can I think whether

[00:46:09] this so the amount of LinkedIn posts

[00:46:11] predicting the downfall of the big four

[00:46:13] is insane. My reaction was let’s wait

[00:46:15] and see because yeah let’s wait and see.

[00:46:17] I’m not old enough to remember but I’ve

[00:46:20] read in books about this when the

[00:46:22] spreadsheet software came in

[00:46:24] consultancies were facing the same issue

[00:46:26] because they have built up large

[00:46:27] personnel of calculator used to be a

[00:46:29] human job right um of people who did

[00:46:31] spreadsheets and the software took away

[00:46:33] that need the client built that

[00:46:34] capability in house for a bunch of

[00:46:36] dollars to to be wired to Microsoft so

[00:46:38] that went away so it’s a similar

[00:46:39] scenario the big guys can probably

[00:46:40] weather the storm they’re already making

[00:46:41] motions right to to deal with it and the

[00:46:43] layoffs as hurtful as they are for the

[00:46:46] impacted people what they mean on the

[00:46:48] business level is the business is

[00:46:49] already responding and they are in the

[00:46:50] luxury position of being able to live

[00:46:52] let people go and still have a business

[00:46:54] as a boutique owner you can probably not

[00:46:55] lay off 20% of your people and still be

[00:46:57] okay I guess right other than you if

[00:46:59] you’re already in crisis which I hope

[00:47:00] you’re not the small guys we like we

[00:47:02] said they can maybe benefit from AI um

[00:47:05] they will have a challenge we we talked

[00:47:06] about the cost of those systems a bit it

[00:47:08] it comes down to investment consulting

[00:47:10] is a cash flow is has strong cash flows

[00:47:13] but they are notoriously not too keen on

[00:47:16] investing significant capex AI I think

[00:47:19] changes that you can see people brag

[00:47:21] posting about their $120,000 a month

[00:47:24] cloud code bills now I think that so

[00:47:26] this notion of I need to have put money

[00:47:28] aside to train my people in AI maybe set

[00:47:31] up some hardware to have some local

[00:47:33] models running pay extra for these for

[00:47:36] the tokens like that that comes to all

[00:47:38] firms of all sizes and that might be a

[00:47:40] bit of a challenge for the smaller guys

[00:47:42] as well but then truth be told most of

[00:47:43] them I encounter do have healthy margins

[00:47:46] so they can probably stomach that the

[00:47:48] guys in the middle are in trouble

[00:47:50] because that’s a bit the nowhere land

[00:47:52] right between you’re bigger than the

[00:47:55] small guys. So you have started to win

[00:47:58] some of this capacity and leverage work

[00:48:00] which is so high highly susceptible to

[00:48:02] the computer doing it but you’re not big

[00:48:04] enough to build the supporting

[00:48:06] infrastructure. If you look again at the

[00:48:07] at the large firms we talked about the

[00:48:09] problems that capability selling creates

[00:48:12] for the business. The big guys don’t

[00:48:14] have these problems not because they

[00:48:16] don’t incur them all the time. They do,

[00:48:17] but because they have invested in

[00:48:19] infrastructure to resolve them. If I go

[00:48:21] to market capability selling, I put the

[00:48:23] client in control of the engagement.

[00:48:24] That typically takes the form of I

[00:48:26] respond to RFPs. So, a lot of work

[00:48:28] nobody pays me for. Well, all the big

[00:48:30] guys have business development

[00:48:32] departments of people who do nothing but

[00:48:34] like they have solved this problem by

[00:48:35] building infrastructure to solve it. The

[00:48:37] middle guys don’t have that. So, they

[00:48:39] are hit from all angles. They need to

[00:48:40] make the investments. they are not big

[00:48:43] enough to have the infrastructure to

[00:48:44] sort of crank the handle and try to win

[00:48:46] more of the old school business, right?

[00:48:48] To to to sort of fill the pockets and

[00:48:51] the the core work they are already doing

[00:48:52] today is highly susceptible to the AI

[00:48:54] stuff. So for for those guys it will be

[00:48:55] most difficult and I think they need to

[00:48:57] be moving fastest and with the most

[00:48:59] decisiveness towards a more for them

[00:49:01] it’s almost like a pivot probably in

[00:49:03] many cases to a more propositionled

[00:49:05] model. Yeah. And also, you know,

[00:49:07] Florian, if the if the junior analyst

[00:49:09] pipeline is shrinking because AI does

[00:49:11] most of that work, what does the next

[00:49:13] generation consultant actually look like

[00:49:16] to you and who is responsible for

[00:49:18] building them, training them?

[00:49:20] >> Oh, it I love discussing that topic. So,

[00:49:21] so if you are firing all your juniors

[00:49:24] because AI does it, you are, if I can

[00:49:26] say it bluntly, stupid, frankly. And

[00:49:28] which I I don’t think anybody’s really

[00:49:31] doing it. I think the way this gets

[00:49:32] reported in the press, oh so and so

[00:49:34] fired all the people because AI no that

[00:49:36] is a blend of structural reasons of

[00:49:38] overhiring uh softening demand in

[00:49:41] sectors and companies just responding. I

[00:49:44] think consultancies are still

[00:49:45] responsible for building their own

[00:49:46] pipeline. I would certainly do that if I

[00:49:48] were to still still was in the business

[00:49:50] of doing that. We’re not we have no

[00:49:51] appetite for running teams anymore here

[00:49:53] in our business. But um the nature of

[00:49:55] those junior jobs I could see them

[00:49:57] changing and honestly we talked about

[00:49:59] where do the prices go. I think what

[00:50:01] technology always does every all types

[00:50:03] of automation it deskills the work to a

[00:50:06] certain extent right so so steel working

[00:50:08] in the medieval times meant you had to

[00:50:10] learn to be a blacksmith which was like

[00:50:12] a 5-year training and then some once

[00:50:14] industrial revolution hits you can work

[00:50:17] with steel by cranking a lever right

[00:50:18] because you operate a machine historians

[00:50:21] don’t come at me for this simplification

[00:50:23] but the same thing could happen to

[00:50:25] knowledge work to a certain extent so

[00:50:26] you could so maybe what what I mean by

[00:50:28] saying that is maybe you go from get an

[00:50:31] excellent degree from a great university

[00:50:32] to come and be our junior analyst for

[00:50:35] two years. That’s the entry role. Get a

[00:50:37] good enough degree and have some

[00:50:39] experience working with LLM systems to

[00:50:42] be our AI supervisor traininee for two

[00:50:46] years. So, so it the structure could

[00:50:48] change in terms of more it becomes more

[00:50:50] an apprenticeship. I teach you how to

[00:50:52] work with the machine rather than an

[00:50:54] artisan apprenticeship like I teach you

[00:50:56] the entire craft, right? And that could

[00:50:58] also mean that salaries come down,

[00:50:59] right? In terms of being a junior

[00:51:01] consultant for, I don’t know, £60,000 a

[00:51:04] year in London, maybe you are uh work

[00:51:06] with the chatbot assistant. I don’t know

[00:51:08] what the job title is going to be. A

[00:51:10] quality assurance person for 40,000. I I

[00:51:13] don’t know. I don’t I don’t know salary

[00:51:15] brands anymore, but I could see that

[00:51:17] happening. Um but then on a broader

[00:51:19] economic perspective, so bad for the

[00:51:20] people who join, but on a broader

[00:51:22] because it’s less salary, but on an

[00:51:23] economic perspective, maybe there’s more

[00:51:25] open positions because of the speed with

[00:51:27] which Claude can chop out 90% stuff that

[00:51:30] needs some quality massaging. I can do

[00:51:33] three times as much, right? So I need

[00:51:34] three times the people on on that role.

[00:51:36] So that is my careful as always. You

[00:51:38] always look like a fool if you make

[00:51:39] predictions. I might look like that in a

[00:51:41] few years. But instinctively, I think

[00:51:43] this is where I could see it going in

[00:51:44] the near term. And and my call to action

[00:51:46] or call to attention for consulting

[00:51:48] owners is if you don’t bring up your own

[00:51:51] uh next level of mid-level senior

[00:51:53] people, your mistake. By the way, not to

[00:51:55] drone on about the proposition, think

[00:51:57] too much, but the more you are

[00:51:59] propositionled and then hopefully

[00:52:01] because of that differentiated and an

[00:52:03] authority market, the more specific the

[00:52:05] requirements of against your the work

[00:52:07] your people do should be. That means you

[00:52:10] almost by definition should need your

[00:52:13] own training program and pipeline. Like

[00:52:15] you should not be able to hire a

[00:52:16] business graduate and they and set them

[00:52:18] off running after 6 months or or well

[00:52:21] maybe not 6 months but it should take an

[00:52:23] intensive additional training in the

[00:52:25] ways of your firm before they can even

[00:52:28] think about becoming productive because

[00:52:30] again this this generalistic undefined

[00:52:33] capability middle that’s the part that’s

[00:52:35] under pressure and yeah universities

[00:52:37] will not train graduates in your

[00:52:39] specific methodology. So better don’t

[00:52:42] fire everybody under 30 to replace them

[00:52:44] with a chatbot. I I don’t think that’s

[00:52:45] happening in reality anyways. I think

[00:52:47] that’s a press story. But yeah, don’t

[00:52:49] make that mistake.

[00:52:50] >> There have been a lot of news, you know,

[00:52:51] around the layoffs from big companies.

[00:52:54] >> When there’s a headline, oh, we we laid

[00:52:55] off 10,000 people. My guess would be

[00:52:58] that there are layers of types of work

[00:53:01] in a corporation that are that have

[00:53:04] always been more vulnerable and

[00:53:06] susceptible to disruption. So the the

[00:53:09] senior solution architects and people of

[00:53:11] the knowledge have probably pretty safe.

[00:53:12] Uh someone with an arts degree writing

[00:53:15] technical documentation for the software

[00:53:17] probably always had a bit more

[00:53:19] precarious job in that overall setting.

[00:53:21] And I think of those 10,000 it’s

[00:53:23] probably a lot of those people. And even

[00:53:26] if we read oh so and so many software

[00:53:28] developers, what type of software

[00:53:29] developers were those? Were they really

[00:53:31] caught to your product and very close to

[00:53:33] the center or were that people who maybe

[00:53:36] do some type of tech support and had

[00:53:38] always were always under pressure of

[00:53:40] being outsourced provider switch

[00:53:42] whatever those those are always been

[00:53:44] sort of not the greatest jobs and that

[00:53:46] should be the headline by the way. Why

[00:53:48] are those jobs not great? Could we fix

[00:53:50] that please?

[00:53:51] >> Um not oh the computer can now do

[00:53:53] software engineering. This is why we’re

[00:53:55] letting everybody go. No, we have always

[00:53:57] these people were always unfortunate and

[00:53:59] at the margins and I think that’s a

[00:54:00] structural issue and this gets masked to

[00:54:03] a certain extent by now oh AI is the new

[00:54:06] game in town and and so forth by the way

[00:54:08] a more positive version of this and a

[00:54:10] bit more provocative one if I can add

[00:54:12] this here I think sometimes what also

[00:54:14] happens is AI just shines their ruthless

[00:54:16] light on nonsense like corporations have

[00:54:19] gotten quite fatty during the corona

[00:54:22] years for a bit it seemed to me and

[00:54:24] previously they’ve got a lot of fat in

[00:54:26] terms of paying for stuff where the

[00:54:28] value is not exactly clear. And now that

[00:54:30] AI comes in and says promises that maybe

[00:54:33] just maybe there’s a different way now

[00:54:35] of doing things. I’m not saying the AI

[00:54:37] can really do it. It just makes this

[00:54:39] promise. That promise that hype existing

[00:54:41] alone is enough for people to turn

[00:54:43] around and look at the processes and

[00:54:44] question the value. And this is where it

[00:54:47] gets interesting because I think

[00:54:48] consulting isn’t free from this. There’s

[00:54:49] many things in consulting that never

[00:54:51] were that valuable to begin with. And

[00:54:53] now that someone says, “Oh, this never

[00:54:55] was I’ve built an AI to do it.” People

[00:54:58] smarter people ask the question, “Yeah,

[00:55:00] should we be buying the AI or could we?”

[00:55:02] Is the fact that someone in their garage

[00:55:05] over two weekends coded a slide builder

[00:55:07] that uses three LLMs? Like literally

[00:55:10] patch this together? If it’s that easy

[00:55:12] to replace, shouldn’t we get rid of the

[00:55:14] process overall? Like, like what’s like

[00:55:16] the presentation right is a bad example,

[00:55:18] but it’s funny to me. We stopped using

[00:55:21] AI notetakers in video conferencing and

[00:55:23] calls a because of the governance. We

[00:55:25] discussed this before. Many corporate

[00:55:26] clients will not allow you to do this

[00:55:28] and I think they have good reason to not

[00:55:29] allow you. Um but the second thing is

[00:55:31] you can take a look into the analytics

[00:55:32] of these things, right? And you can get

[00:55:35] very astonishingly low numbers for who

[00:55:38] actually ever clicks and reads them

[00:55:39] which tells you something about having

[00:55:41] those notes in the first place which

[00:55:43] then shines a light on the junior who

[00:55:46] was in the call to take notes and build

[00:55:47] by the hour in our consulting past. So I

[00:55:50] think that effect is also real. The AI

[00:55:52] reset forces companies to reflect on do

[00:55:55] we even need this in the first place?

[00:55:57] And there are cases, there are roles,

[00:56:00] there are jobs, there are types of work

[00:56:01] where the answer, unfortunately, is

[00:56:03] honestly no. So, and then it doesn’t get

[00:56:06] replaced by I, it just gets cut

[00:56:08] completely and that’s that. And again,

[00:56:10] that’s not what you read in the press,

[00:56:11] of course.

[00:56:12] >> Yeah. Yeah. So, moving towards the last

[00:56:15] two questions of our conversation. So,

[00:56:17] if you had to give one piece of advice

[00:56:19] to consulting leaders who are actively

[00:56:22] resisting AI adoption at the moment,

[00:56:25] what would you say to them? uh do they

[00:56:27] exist? So, so I think you might not be

[00:56:29] wrong. I want to acknowledge I can see

[00:56:31] reasons for why to resist this for what?

[00:56:33] So, um you might not be wrong, but the

[00:56:36] times seem to be moving in a different

[00:56:38] direction. Uh I recently made this

[00:56:40] comparison from AI to the to the car to

[00:56:44] the automobile. Uh in the quote unquote

[00:56:47] western world post World War II, a lot

[00:56:49] of people didn’t like cars because they

[00:56:51] were rich people’s toys. They cause a

[00:56:54] lot of crashes. Like I mean Germany we

[00:56:56] had Germany exports tons of cars. We

[00:56:58] used to have at the turn of the century

[00:57:00] protest movements against cars with

[00:57:03] posters you can see in museums today

[00:57:04] where they are called uh children the

[00:57:07] killers of children. That’s what cars

[00:57:08] used to be. So there were very many good

[00:57:11] reasons to say as a means of

[00:57:13] transportation from a societal

[00:57:15] perspective the car is not a very smart

[00:57:18] option even in terms of economics and

[00:57:21] still that’s where it went because the

[00:57:23] the powered interests and the money and

[00:57:25] the economic circumstance and having to

[00:57:28] build a bunch of infrastructure after

[00:57:29] lots of circumstantial pressures made it

[00:57:32] so so your resistance against AI might

[00:57:34] be entirely correct and sound and you

[00:57:36] might have all the right reasons the

[00:57:38] times might still turn against you. So,

[00:57:40] just be aware of that. America still

[00:57:41] runs on cars, not trains. Even though

[00:57:43] there’s many, many good factual reasons

[00:57:45] why they should have more trains, it’s

[00:57:47] just not the reality.

[00:57:48] >> So, if you don’t want to go the way of

[00:57:49] trains, um maybe find a way dealing with

[00:57:52] AI and a good way of to not think about

[00:57:55] it at all. But sorry for tutting my own

[00:57:57] horn here too much. But if you can shift

[00:57:59] your business to a more propositionled

[00:58:01] stance and often you already know the

[00:58:04] problems you solve. You’re just not

[00:58:05] leading with them in terms of your go to

[00:58:07] market and messaging and way of

[00:58:08] thinking. You bury them in the case

[00:58:10] study. You say we know Salesforce here

[00:58:12] needs Salesforce yada yada yada. Look at

[00:58:13] this problem we’ve solved here this

[00:58:15] problems. So if you can just flip the

[00:58:17] order and start thinking about and

[00:58:19] talking about your client’s problems and

[00:58:21] how you solve them first, you are

[00:58:23] already improving the entire business of

[00:58:25] your consulty. If you do that enough,

[00:58:28] you might eventually find yourself in a

[00:58:30] position where now clarity and

[00:58:33] stringency have increased, complexity

[00:58:35] has come down, business processes have

[00:58:38] hardened to an extent which way you

[00:58:40] could say, hey, maybe we could save an

[00:58:41] additional six hours if a computer could

[00:58:43] help here. And then that AI can come in

[00:58:46] and naturally support the problem

[00:58:48] solving process that you are selling and

[00:58:51] where you take your pride in and which

[00:58:53] is your way of actually helping people

[00:58:55] and that is completely different than we

[00:58:57] are an AI first consultancy helping you

[00:58:59] transform your business of AI which is

[00:59:01] nothing I run a business I don’t want

[00:59:03] anyone to transform it that sounds like

[00:59:06] way too much work and risk nobody’s

[00:59:08] buying that stuff so ignore that noise

[00:59:09] in the market focus on the problems you

[00:59:11] solve and once you are robust trust and

[00:59:14] successful enough in bringing in the

[00:59:15] same problem ever again deepening

[00:59:17] expertise building an authority for the

[00:59:19] in your market for being good at solving

[00:59:20] that then you can ask the question could

[00:59:22] AI help me do this that’s one thing the

[00:59:25] second question the second answer we

[00:59:26] touched upon this earlier if you are in

[00:59:27] IT consulting whether you like AI or not

[00:59:30] it is making the lives of your clients

[00:59:32] more complicated because they get the

[00:59:34] same questions from their bosses and it

[00:59:36] is your job as a consultant to help them

[00:59:38] navigate that so AI as a complication as

[00:59:42] noise in their day-to-day help them so

[00:59:45] that um it’s just part of your it’s just

[00:59:47] become part of your job description much

[00:59:48] like any other IT disruption would be

[00:59:50] >> after years of working inside consulting

[00:59:52] firms through exactly this kind of you

[00:59:55] know disruption and changes what’s the

[00:59:58] one lesson you had to learn the hard way

[01:00:00] >> the stickiness of the old ways brutal

[01:00:04] and everything I say seems the Americans

[01:00:07] have a saying it’s easy from far far

[01:00:09] from easy because the conceptual logic

[01:00:12] of hey just talk about the problem first

[01:00:14] and structure your go-to marketing

[01:00:15] message that you can do that in an

[01:00:17] afternoon to your earlier point with

[01:00:19] chat GPT you can probably do it in 15

[01:00:22] minutes getting your entire organization

[01:00:24] of 30 consultants who for their entire

[01:00:26] existence have been trained in doing the

[01:00:29] opposite are now working in probably a

[01:00:31] set of incentive structures geared

[01:00:33] towards the opposite and dealing with

[01:00:35] clients for their entire life have been

[01:00:37] trained in buying the opposite just

[01:00:40] because there was no better alternative

[01:00:41] is the hard part. So coming back to

[01:00:43] something earlier, you asked me how does

[01:00:45] this translate to the people’s side of

[01:00:47] consulting getting the proposition work

[01:00:50] and the ideas and the structures is you

[01:00:53] can get to a maybe decent enough version

[01:00:56] of that with an AI tool pretty quickly.

[01:00:58] That’s not the hard part. Organizational

[01:01:00] change, the making it stick, the dealing

[01:01:03] with exceptions, getting some clarity in

[01:01:05] the messiness that still persists

[01:01:06] because everybody involved is a person.

[01:01:08] That is the hard part and that never

[01:01:11] goes away which is good news because

[01:01:13] that is also where the market for

[01:01:14] consulting has always been and will

[01:01:17] continue to be.

[01:01:18] >> Well Floren thank you so much for your

[01:01:21] grounded and insightful conversation and

[01:01:23] sharing such such great insights with

[01:01:25] our listeners today.

[01:01:26] >> Thank you for hosting me. It was it was

[01:01:28] a lot of fun and continues to be for me

[01:01:29] to talk about this stuff and and great

[01:01:31] questions. I really appreciate the

[01:01:32] discussion. Thank you

[01:01:33] >> to everyone listening. Thanks for

[01:01:35] joining us on Tech Unhinged. Until next

[01:01:37] time.