[00:00:00] [music]
[00:00:04] Welcome back to another episode of Tech
[00:00:06] Unhinged powered by code district where
[00:00:08] tech gets human. I’m your host Rabia
[00:00:10] Javeed and joining me today is Lydia who
[00:00:12] is the chief information officer at
[00:00:14] appreciation financial co-founder of
[00:00:16] Genie AI and founder of the people’s
[00:00:18] accords with over eight years in data
[00:00:21] and AI strategy and financial services.
[00:00:23] She has seen exactly what it takes to
[00:00:25] modernize legacy industries with AI.
[00:00:28] Welcome to Techan, Lydia.
[00:00:29] >> Thank you. Thank you so much for having
[00:00:31] me. I’m really excited to be here with
[00:00:33] you today, Labia.
[00:00:34] >> We’ll start with the ice breakers,
[00:00:35] Lydia. We see that most CIOS have a very
[00:00:38] linear path, but yours also has a
[00:00:40] martial arts chapter in it. How did that
[00:00:43] happen for you?
[00:00:44] >> So, it happened. I mean, I got my
[00:00:45] master’s degree. I had kind of achieved
[00:00:47] a lot of things I wanted to achieve
[00:00:49] before. Let’s say I was, you know, going
[00:00:52] to get married and have kids. I I wanted
[00:00:53] to study martial arts. I had a a
[00:00:56] master’s degree. I had worked at the UN.
[00:00:58] I had experience doing research for the
[00:01:00] World Bank. A lot of things that were
[00:01:01] really exciting to me, but I wasn’t
[00:01:03] quite ready to just go to an office
[00:01:05] environment and this was before co so
[00:01:07] working remotely was not as common at
[00:01:10] that time. And in my company, I was able
[00:01:12] to work remotely and study martial arts
[00:01:14] as well as do tech things. I kind of
[00:01:16] worked in the background and I slowly
[00:01:19] automated my job. I made a lot of little
[00:01:21] bots that could do my job for me and
[00:01:23] essentially um enhanced our data
[00:01:25] analytics at the company and eventually
[00:01:27] it became so uh noticeable that the CEO
[00:01:30] called me and was like you need to come
[00:01:32] and like show me what you’re doing and I
[00:01:34] presented to him some of the programs I
[00:01:36] had worked on and over a few years I was
[00:01:39] promoted to CIO and so that was a really
[00:01:41] exciting pathway to know that it wasn’t
[00:01:43] just something I don’t know got handed
[00:01:46] to me I built it out of my love for data
[00:01:49] and my love for techn technology and uh
[00:01:51] it was an it was an excellent
[00:01:52] opportunity to combine my skill set as
[00:01:54] well as showing the power of the
[00:01:56] internet and as we’ll talk about later
[00:01:58] throughout this podcast what can AI do
[00:02:00] there’s you know the internet has opened
[00:02:02] up information to everyone everywhere
[00:02:04] but AI is in some ways the next level of
[00:02:06] that so um you know thank you for your
[00:02:09] interest in in my background and for
[00:02:10] having me here today
[00:02:11] >> yeah I know we say that where there is a
[00:02:13] will there is a way we see that 92% of
[00:02:16] leaders say AI is critical but fewer
[00:02:18] than 25% have systems ready to support
[00:02:20] it. That gap has now a name we call the
[00:02:23] modernization paradox. How do you define
[00:02:25] it in plain terms?
[00:02:26] >> I think two words really like context
[00:02:29] and junk data summarize why a lot of
[00:02:31] people are they want AI. They know what
[00:02:33] it is. They see that there’s value
[00:02:35] there, but it’s very hard to put it into
[00:02:37] practice because you have a lot of, as
[00:02:39] you mentioned, the legacy systems, but
[00:02:42] also people, not that people are legacy
[00:02:44] or outdated, but how do we have a we
[00:02:46] have a massive workforce that of all
[00:02:49] ages and all backgrounds and in order to
[00:02:52] actually put this in practice, it would
[00:02:54] need to cover a lot of different
[00:02:55] knowledge gaps and levels. So I think
[00:02:57] one of the first steps would actually be
[00:02:59] not necessarily modernizing the
[00:03:01] technology but helping the workforce
[00:03:03] understand how AI can work in their role
[00:03:06] and actually educating them on using it
[00:03:08] in their daily life before it can even
[00:03:10] become fully agentic because engineers
[00:03:12] can only do so much. An engineer knows
[00:03:15] his job. He knows code. He knows
[00:03:17] systems. He knows architecture but he
[00:03:19] doesn’t know what the person in human
[00:03:20] resources is doing. He doesn’t know how
[00:03:22] to tell her how to do her job better.
[00:03:25] And so in order for AI to work across
[00:03:27] departments, we really need to give it
[00:03:28] to the people using it and then let them
[00:03:31] define how it will be used and
[00:03:32] eventually the system will modernize and
[00:03:34] adapt once the people are ready.
[00:03:35] >> Yeah, know that makes a lot of sense. So
[00:03:37] Lydia, when 60 to 80% of IT budgets are
[00:03:40] just keeping old systems alive and
[00:03:42] technical debt is sitting at around 1.52
[00:03:46] trillion, what does it actually take for
[00:03:48] leadership to see modernization as a
[00:03:50] revenue decision? that is I so I think
[00:03:53] it’s a it’s a bit of a paradox type of
[00:03:55] question because modernization at least
[00:03:58] I think we’re starting to see has its
[00:04:00] own risks in terms of once we modernize
[00:04:03] away from these systems or we’re fully
[00:04:05] on cloud computing or we’re fully
[00:04:07] agentic AI and we’re using it to do all
[00:04:09] of our tests the consumption rate of
[00:04:11] data increases and then the storage
[00:04:13] costs of the data increase and so in
[00:04:16] some ways our modernization costs can
[00:04:18] continually increase and also stress the
[00:04:20] bottom line of the business. However, we
[00:04:22] probably need to focus on balance. We
[00:04:24] need to have our old legacy systems,
[00:04:26] maybe especially for highly regulated
[00:04:28] industries where you must store data for
[00:04:30] very long periods of time. Find someone
[00:04:33] who can put all of that old data that
[00:04:35] just needs to be stored for legal
[00:04:36] purposes and more cost-effective
[00:04:38] systems. And then for fast-paced uh high
[00:04:41] context action items that can go in the
[00:04:43] cloud, that should be easy to access. It
[00:04:45] should be able to interact with AI
[00:04:47] systems much more quickly than the old
[00:04:49] legacy environments. I think there’s a
[00:04:51] bit of a play between what is legally
[00:04:54] required of companies and then what is
[00:04:56] needed to modernize and adapt and like
[00:04:58] adopt AI across departments. Another
[00:05:01] thing that I’ve noticed is and I think
[00:05:03] we’ll talk about a bit more is just junk
[00:05:05] data. Like how much of the data that
[00:05:07] we’re storing is actually required by
[00:05:09] law or it’s just something that is junk
[00:05:12] and needs to be deleted like on
[00:05:13] individual drives or in team drives from
[00:05:16] old projects and files that’s just
[00:05:18] stored everywhere. So, I’m not sure what
[00:05:20] it would take for leadership to see it
[00:05:22] that way. Probably somebody going to the
[00:05:24] CFO with some written reports of here’s
[00:05:26] exactly how much we’re spending on
[00:05:28] storage. Here’s how much we’re spending
[00:05:29] on development costs and here’s how much
[00:05:31] it’s going to cost to clean it up and
[00:05:33] model different pathways for storage.
[00:05:35] So, it’ll cost this much to store here
[00:05:37] to clean it. And once the CFO can
[00:05:40] clearly see those cost comparisons, it
[00:05:42] will it will kind of get to the CEO and
[00:05:44] and change will will start to happen.
[00:05:46] >> Yeah. No, absolutely. Lydia, we see that
[00:05:48] Mackenzie’s 2025 survey found that
[00:05:51] twothirds of companies are using AI
[00:05:53] across multiple functions, but only a
[00:05:55] third of them are able to scale it
[00:05:57] enterprisewide. What’s creating that
[00:06:00] gap?
[00:06:00] >> I think it’s all comes down to data.
[00:06:02] Sometimes data and the people. So
[00:06:04] there’s many factors obviously. I mean
[00:06:06] anyone who works with AI or in a company
[00:06:09] will talk about security risks, token
[00:06:11] bloat, AI hallucinations. kind of
[00:06:14] provide accurate information in context
[00:06:16] to enable for especially larger
[00:06:18] companies to be able to use it. I’ve
[00:06:20] I’ve talked to many people from my own
[00:06:23] network as well as within our company
[00:06:25] about why aren’t you using AI and many
[00:06:27] people just don’t feel that it’s
[00:06:29] relevant to them until somebody takes
[00:06:32] their hand and sits down with them and
[00:06:34] and shows them ask this ask that here’s
[00:06:37] and then it has to be a realization
[00:06:39] process from the individual users of
[00:06:41] what it can do for their job in order
[00:06:43] for adoption to start taking place and
[00:06:45] once you have a larger base of people
[00:06:47] inside of a company willing to adopt the
[00:06:49] technology then more use cases I think
[00:06:52] will start being undertaken throughout
[00:06:54] all of the different departments. So
[00:06:56] engineering is usually pretty quick to
[00:06:58] want to use it but once they apply it
[00:07:00] the end users might not have any they
[00:07:02] might not have any idea what to do with
[00:07:04] this thing you’ve just brought to them
[00:07:05] and so it kind of just dies. Uh
[00:07:08] unfortunately I’ve seen that happen a
[00:07:10] lot. I think some of the resistance is
[00:07:11] that is this here to replace me? I don’t
[00:07:14] want to do this. I’m skeptical of this.
[00:07:15] this is, you know, this is going to take
[00:07:17] my job and to kind of show people that
[00:07:20] what it can do for you because there’s
[00:07:22] everyone has self-interest, right? And
[00:07:23] so if my self-interest is to avoid this
[00:07:26] new change, to avoid this technology,
[00:07:28] then it’s harder to get the organization
[00:07:30] to actually adopt it enterprisewide.
[00:07:32] Whereas if the individuals throughout
[00:07:34] each department see their self-interest
[00:07:36] in using it not just to keep the job but
[00:07:39] to make job easier to make the job
[00:07:41] faster, better to have quicker quality
[00:07:43] of reports. Maybe someone hates doing
[00:07:45] PowerPoints and they don’t realize if I
[00:07:47] use claude I can get my PowerPoint done
[00:07:48] in 30 seconds and just update it now in
[00:07:51] Google Slides. And I like you said I
[00:07:53] think training is really important but
[00:07:55] then sometimes and I’ve seen this at
[00:07:56] large companies as well. I won’t name
[00:07:58] drop. They’ll take some random person
[00:08:00] and say, “Okay, you’re in charge of AI.”
[00:08:02] And that person has no idea how to even
[00:08:04] use AI themselves. So, it’s it’s just a
[00:08:07] bit funny. Eventually, I think I don’t
[00:08:10] want to say it. Eventually, everyone
[00:08:11] will be using it, but I already have AI
[00:08:13] in my car. I didn’t have a choice. So, I
[00:08:15] think we everyone will eventually be
[00:08:16] using it and exposed to it in some way,
[00:08:18] but that gap will close probably in the
[00:08:20] next 5 years.
[00:08:21] >> You’ve also said that AI is only as good
[00:08:23] as the data beneath it. So Lydia, maybe
[00:08:25] you can um talk tell our listeners about
[00:08:28] that. What does that look like in
[00:08:30] process?
[00:08:31] >> Absolutely. And I just I wanted to touch
[00:08:33] on the the AI agentic AI part as well
[00:08:36] where everyone wants to say this is AI,
[00:08:39] this is AI and some things it’s actually
[00:08:41] not cost effective to put AI especially
[00:08:43] if the data underlying it or the context
[00:08:45] isn’t really there. So much of what we
[00:08:47] call agentic AI can be done by
[00:08:50] traditional modernization by traditional
[00:08:52] automation practices of just using
[00:08:54] Python or Java or appcripts things that
[00:08:57] can be just a simple cost-ffective you
[00:08:59] don’t need billionoint LLM attached to
[00:09:02] get some task done but just calling AI
[00:09:04] has been the exciting thing and AI is I
[00:09:07] would say only as good as the data
[00:09:08] beneath it because if you have a lot of
[00:09:10] junk data that’s just old it’s outdated
[00:09:13] it’s not attached to you know I’ll speak
[00:09:15] in the context of appreciation ation. We
[00:09:17] have a lot of financial services
[00:09:19] representatives who come and go over the
[00:09:21] years. Some people have special, you
[00:09:23] know, relationships with their clients
[00:09:25] who are either their family members or
[00:09:27] their own, you know, it’s their own
[00:09:29] policy or it’s somebody under 18 who
[00:09:32] they’ve written a life insurance policy
[00:09:33] for a child. But if the AI doesn’t have
[00:09:36] the proper context to understand, for
[00:09:38] example, you want to use it for
[00:09:39] marketing. I think copywriting and
[00:09:41] marketing is one of the easiest tasks
[00:09:42] for generative AI to accomplish these
[00:09:44] days. It can still create a lot of
[00:09:46] problems for a salesperson. They want to
[00:09:48] implement AI, but it doesn’t understand
[00:09:50] their data. It doesn’t have the context
[00:09:52] to know, oh, I can’t send an email to
[00:09:54] this person because they’re 12 years old
[00:09:56] or I shouldn’t try and set an
[00:09:58] appointment with that person because
[00:09:59] that’s their mom. They’re they can just
[00:10:01] call their mom. And so, there’s just a
[00:10:03] lot of little missteps that can happen.
[00:10:05] And I’m using a very generic example,
[00:10:07] but I think it’s one that’s easy to
[00:10:09] relate to of those simple pieces of
[00:10:11] information that AI is very quick to
[00:10:13] overlook. And so not only the data
[00:10:15] underlying it, if is it old, are the
[00:10:18] policies, is there a good record if your
[00:10:20] policy is active, surrendered, lapsed,
[00:10:22] does it know all of those things before
[00:10:23] it’s going out in marketing? But also,
[00:10:25] is it able to have the right context of
[00:10:28] all the relationship underlying the
[00:10:29] data? And sometimes, especially at our
[00:10:32] company, things can change quickly. We
[00:10:34] have agents and clients. We have over a
[00:10:36] 100,000 clients all around the country.
[00:10:38] And things can change very quickly from
[00:10:40] day to day, from minute to minute. And
[00:10:42] having an AI that has secure access to
[00:10:44] our database is one thing where we don’t
[00:10:47] necessarily want to give it access to
[00:10:49] every single data point. So now you need
[00:10:51] somebody to manage that context and make
[00:10:53] sure that it’s giving accurate
[00:10:55] responses. And we we didn’t really find
[00:10:57] that our agents actually enjoyed using
[00:11:00] the AI we put in our back office because
[00:11:02] the context was not able to keep pace
[00:11:04] with the change um throughout throughout
[00:11:07] the the company. So that’s why I think
[00:11:09] just managing data is sometimes it takes
[00:11:12] a certain personality. A lot of people
[00:11:14] especially in small companies, big
[00:11:16] companies, doesn’t matter to to look at
[00:11:18] the little details of the data is
[00:11:20] actually playing a very big role in how
[00:11:22] successful the AI is going to be. Lydia,
[00:11:24] what’s your particular framework for
[00:11:26] assessing and making decisions around
[00:11:29] legacy systems, you know, as for the as
[00:11:31] for the exposure or the work that you’ve
[00:11:33] done so far? I love the the word intent
[00:11:36] and also how you know it can mean
[00:11:38] different things for different people
[00:11:39] especially across departments. If I’m in
[00:11:41] the data or technical department I think
[00:11:44] of my intent for one thing for the
[00:11:46] system but if I’m missing the context
[00:11:47] from the other department that’s that
[00:11:49] can be a huge issue later down the road
[00:11:51] where I didn’t know something that I
[00:11:53] should have included and what what was
[00:11:55] behind the scenes. And so when I’m
[00:11:56] handling problems a lot of times I try
[00:11:58] to start and I like to take a layered
[00:12:00] approach. I try to start with what is
[00:12:02] the most necessary thing. What is the
[00:12:04] 80% problem here and I tend I like to
[00:12:07] write lists but I’ll make a list of what
[00:12:09] I see as the most visible I would say
[00:12:12] bottom line critical problems like what
[00:12:14] is going to affect our sales what is
[00:12:16] going to affect the accuracy of our
[00:12:18] information and start peeling it apart.
[00:12:20] So start with that top uh 80% issue of
[00:12:23] there was an automation error in one of
[00:12:25] our CRM for example where clients were
[00:12:28] not being labeled as clients and once
[00:12:31] their policy was issued they were still
[00:12:32] being labeled as leads. So there was a
[00:12:34] marketing error going out to some of
[00:12:36] these clients marketing to them as leads
[00:12:38] when they were already clients. So
[00:12:40] that’s just an example of here’s like a
[00:12:41] number one big level issue that we need
[00:12:43] to tackle right away. as I’m going
[00:12:45] through the details, I can find the
[00:12:47] outliers. I can find the small use cases
[00:12:50] and then as we’re handling the 80%
[00:12:52] issue, we tend to take care of the 20%
[00:12:54] along the way. So I like to take like
[00:12:56] the layered approach but also a human
[00:12:58] approach to consult with the people um
[00:13:00] not just on my team but maybe with other
[00:13:02] teams to see what are they dealing with
[00:13:04] and how is you know my approach here’s
[00:13:07] my solution to the problem but make sure
[00:13:08] I get input from other departments and
[00:13:10] other people who are touching the issue
[00:13:13] as I guess is one way to say it that to
[00:13:15] see how does my solution work with
[00:13:17] theirs for example the finance team
[00:13:20] might have a certain way that they’re
[00:13:21] doing something and that they’re used to
[00:13:23] doing something or that from the
[00:13:25] financial perspective works best. But
[00:13:27] then when I’m looking at the sales team
[00:13:28] and the sales department, my solution is
[00:13:31] more suited for sales than for finance.
[00:13:33] And so I have to approach things in
[00:13:35] layers as well as with that human
[00:13:37] context and I can’t just work in a silo
[00:13:40] of what I think is logically the best
[00:13:41] because one time we implemented a system
[00:13:43] that was in our worldview perfect. It
[00:13:46] was automated. It had all the right
[00:13:48] steps. It was going to be great for the
[00:13:50] agents. And when we actually released it
[00:13:51] to the field, everybody hated it because
[00:13:54] it was perfect for us. It was very
[00:13:56] logical, but it didn’t actually adapt to
[00:13:58] how the agents do business daytoday, how
[00:14:02] they actually adapt to when they need to
[00:14:04] call somebody into a meeting, when they
[00:14:06] need to do things kind of on the fly.
[00:14:08] And so we went back to the old process.
[00:14:10] We thought people hate filling out this
[00:14:12] form. It’s so boring. They have to type
[00:14:14] it. Let’s just automate everything. And
[00:14:16] so my approach has changed over time
[00:14:19] from doing things like purely logically
[00:14:21] like purely by the steps of what’s
[00:14:23] perfect to you know what works for the
[00:14:25] people and to try and not necessarily
[00:14:26] get every single thing perfect right
[00:14:28] away but to take care of the major
[00:14:31] absolutely necessary issues up front and
[00:14:33] then once those are handled we can take
[00:14:35] care of the use cases like deeper to the
[00:14:37] core of the system that will over time
[00:14:40] um make it a more perfect system at
[00:14:41] least for our our database. I mean have
[00:14:43] you had any similar or different
[00:14:45] approaches from some of the other
[00:14:46] leaders that you’ve talked to?
[00:14:48] >> When we talk about legacy modernization
[00:14:50] in particular so there lot lots of
[00:14:52] different frameworks and approaches that
[00:14:54] people are using and then that also
[00:14:56] comes down to you know how big of a
[00:14:58] number that particular business is
[00:15:00] looking at in terms of their headcount
[00:15:02] is it small scale is it an enterprise
[00:15:04] level or midsize perhaps and what’s
[00:15:06] working out for them and what’s not. I
[00:15:08] mean some of them are still juggling
[00:15:10] with their internal hierarchies about
[00:15:12] you know the idea of modernization while
[00:15:14] they are still stuck you know with their
[00:15:16] older systems the new people are still
[00:15:18] trying to convince them on approving the
[00:15:20] budgeting for why this middle layer is
[00:15:22] now important and now it’s high time
[00:15:25] they automate their manual tasks the
[00:15:27] debate is still about modernization
[00:15:29] perhaps but you know the jargon has
[00:15:31] probably changed a bit with aentic AI or
[00:15:33] workflows perhaps because you know yes
[00:15:36] absolutely you know the excel sheet that
[00:15:38] you were taking probably 3 days to get
[00:15:40] up with. Now Claude can probably do it
[00:15:42] in 30 seconds as you mentioned. So there
[00:15:44] are a lot of bottlenecks that we witness
[00:15:46] in the current landscape. They are on
[00:15:48] the good side and on the bad side. But
[00:15:51] you know there’s definitely progress
[00:15:52] too. But again the ones who are trying
[00:15:55] to catch up or in the game I think
[00:15:57] things are probably on the brighter side
[00:15:58] for them. But you know you’re also
[00:16:01] hearing about the massive layoffs at
[00:16:04] different verticals across the world you
[00:16:06] know and
[00:16:08] >> it is it is quite evident that if you do
[00:16:11] not have the right skill set to be able
[00:16:13] to use AI as your assistant you will be
[00:16:17] replaced you know I mean to be able to
[00:16:19] have an assistant you need to have a
[00:16:20] good grip on your skill set first and if
[00:16:24] you don’t have it then 100% you know
[00:16:26] you’re no more in the game
[00:16:27] >> and I think honestly people. So when we
[00:16:30] talk about problem solving and then you
[00:16:32] know going into just getting stuck where
[00:16:35] some things get stuck the individual
[00:16:37] matters so much like I’ve noticed one of
[00:16:39] the biggest bottlenecks I’ve come across
[00:16:41] is people not wanting you to look over
[00:16:43] their shoulder feeling judged because
[00:16:45] sometimes if you’re coming from the
[00:16:47] technical department and then going into
[00:16:49] the like non-technical department they
[00:16:51] feel like are you saying you’re better
[00:16:52] than me? Are you saying you’re smarter
[00:16:54] than me? When really I’m trying very
[00:16:55] hard to say I want to help you. I want
[00:16:57] to help you. and and but when they hear
[00:16:59] me say I want to help you be faster and
[00:17:00] and do things in a more efficient way to
[00:17:03] me that’s an exciting thing to them
[00:17:05] that’s like well what’s wrong with me
[00:17:06] like are you saying I’m not good and so
[00:17:08] the personality issue has been really
[00:17:10] interesting of like how to not make
[00:17:12] people feel like they’re not good enough
[00:17:14] because they see AI as telling them
[00:17:16] they’re not good enough and if they
[00:17:17] don’t come from a technical background
[00:17:19] they’re bit it’s a bit intimidating to
[00:17:21] say well here just go click on this
[00:17:23] Google studio over here you can create
[00:17:25] your own workflow it’ll do everything
[00:17:26] for you and I I think it’s just a matter
[00:17:28] of the effort coming from the different
[00:17:30] department heads because they already
[00:17:32] have a relationship with their team. So
[00:17:34] if the department head has some idea of
[00:17:36] what to do, they can go to their team
[00:17:38] and show them what to do and then they
[00:17:40] will, you know, essentially have a
[00:17:42] better more secure future. But I’ve also
[00:17:44] noticed the engineers are the ones
[00:17:46] getting laid off too. So you know, no
[00:17:48] one is truly safe uh from from the whole
[00:17:51] AI revolution. But the more you know,
[00:17:54] the more that we try to adopt it into
[00:17:56] our daily routines, I think the company
[00:17:58] will be better. And also job security.
[00:18:00] >> Most experts agree that you shouldn’t
[00:18:02] rip out a legacy system all at once. The
[00:18:04] safer move is running the old and new
[00:18:07] system side by side until you’re
[00:18:09] confident enough to cut over. In
[00:18:11] practice, what does it look like to you?
[00:18:13] >> So, I absolutely agree. Changing too
[00:18:15] many things at once has is horrible. And
[00:18:17] I I tend to save old code at least just
[00:18:20] in case I ever need to put it back if
[00:18:22] something happens. We like to test
[00:18:23] things rigorously before deploying them
[00:18:26] and making small changes at a time
[00:18:28] instead of pushing massive updates
[00:18:29] because some things that even the
[00:18:31] developers forgot are connected to like
[00:18:34] some reports are connected to something
[00:18:35] you push an update and now like
[00:18:37] something over here is broken that
[00:18:38] nobody noticed or nobody saw. And it’s
[00:18:40] amazing how often that happens. And I
[00:18:43] would say, you know, it’s not just at my
[00:18:44] company. We are a smaller company in
[00:18:46] terms of what our, you know, our home
[00:18:48] office size is. We have a large
[00:18:49] distribution. We’re actually one of the
[00:18:51] we are like the number one company and
[00:18:54] our niche for serving teachers and in
[00:18:56] the 403b market, but it’s still a very
[00:18:58] niche market. I mean, we’re doing over
[00:19:00] $300 million in premium, I think, a
[00:19:03] month some and sometimes for um just
[00:19:05] for, you know, rollovers and life
[00:19:07] insurance and flow business, but still
[00:19:09] we’re considered like a a smaller
[00:19:11] company. And so we do our best to take
[00:19:14] our time and test things and but then
[00:19:15] sometimes we just have to like push
[00:19:17] stuff out and then fix it fix it as it
[00:19:19] comes out. One of the reasons that my
[00:19:20] role became really important in our
[00:19:22] company as we grew because we started in
[00:19:24] the founders’s garage with like you know
[00:19:26] group of five people. They started their
[00:19:28] company and they grew it to one of the
[00:19:30] to the biggest company in our niche but
[00:19:32] over time that changed our technology
[00:19:34] dramatically. Like we didn’t used to
[00:19:36] have any technology. It was like word
[00:19:37] documents and stuff and now we have a
[00:19:39] whole back office. We have like all
[00:19:41] kinds of platforms and automations and
[00:19:43] and so those requirements communicating
[00:19:46] the requirements to the developers was
[00:19:48] one of the consistent reasons for
[00:19:50] failure in deployments because you would
[00:19:53] have somebody who would just say hey we
[00:19:55] want this we need it now and then it
[00:19:57] just gets deployed and it wasn’t tested
[00:19:59] because there was such a fast-paced
[00:20:00] environment. But then as we go back and
[00:20:02] as we’ve grown to mature, we now make
[00:20:05] sure that we have testing, we have
[00:20:06] implementation, we have a protocol, we
[00:20:08] have support tickets. And sometimes for
[00:20:10] the people who are used to the good old
[00:20:12] days, it’s it can be a bit this tug and
[00:20:14] pull between wanting to move quickly and
[00:20:16] wanting to get things out there and also
[00:20:18] needing to wait for, hey, we have to
[00:20:20] test it, we have to iterate it, we have
[00:20:21] to fix this. And so there’s, you know,
[00:20:23] we still feel some of that tug and pull
[00:20:25] between the, you know, the wanting to
[00:20:27] get things done quickly and out there.
[00:20:29] And that was something that I have
[00:20:31] really pushed forward very strongly in
[00:20:33] my role was I remember there was a
[00:20:34] program one of the largest programs that
[00:20:36] I manage that’s it’s several million
[00:20:38] dollars of incentive programs that we do
[00:20:40] to help our agents build their business.
[00:20:42] I built this program from the ground up
[00:20:44] and now it’s one of the defining things
[00:20:46] that attracts people to our company. But
[00:20:48] it was essentially first brought to me
[00:20:49] as an idea on like a Tuesday and then
[00:20:52] they’re like so can you roll this out on
[00:20:54] Thursday? And you know, that’s the
[00:20:55] fast-paced environment of also having
[00:20:57] sometimes a smaller company where people
[00:20:59] get to meet the CEO, people know him,
[00:21:01] and they see how vibrant his personality
[00:21:03] is. But my role became very important as
[00:21:06] we grew to put some of that structure in
[00:21:08] and say, “Well, I can get you an outline
[00:21:10] by Thursday and some general and a test
[00:21:12] group of how it’s going to work.” And
[00:21:13] then we tested it and we iterated it.
[00:21:15] And now I have 5 years of ROI history to
[00:21:18] show how we’ve developed the program,
[00:21:19] how we’ve iterated the program, and how
[00:21:21] we’ve now made it into something that an
[00:21:23] agent, we can see when they come in, are
[00:21:25] they going to be successful or not, and
[00:21:26] we can support them into now making
[00:21:28] $50,000 a month, which is really cool. I
[00:21:30] think this is really, really impressive.
[00:21:32] And you know, you being this open about
[00:21:34] your entire journey, I think this is
[00:21:36] also very inspiring for a lot of the
[00:21:37] people out there. And I think that one
[00:21:39] thing from your conversation that I’ve
[00:21:41] gotten is that how you’ve always put
[00:21:43] yourself and stepped out of your comfort
[00:21:45] zone and you know things and change only
[00:21:47] happens when you know you’re not in your
[00:21:50] comfort zone and you’re always just you
[00:21:51] know moving taking certain actions with
[00:21:53] certain risk analysis and then just
[00:21:55] experimenting and learning perhaps every
[00:21:58] day. So yeah amazing journey. Yeah, and
[00:22:01] I I’m I know that I’m the I’m a very
[00:22:04] non-traditional CIO and I’m proud of it.
[00:22:07] You know, I’m I’m absolutely comfortable
[00:22:09] with where I’m at and I love to learn
[00:22:10] more. I just did a new in January I did
[00:22:12] an MIT professional education
[00:22:14] certificate because I want to I know a
[00:22:17] lot about AI, but I want to have the
[00:22:19] certificate behind me to back me up to
[00:22:21] show that hey, I do know what I’m
[00:22:23] talking about. And it was funny as I was
[00:22:24] going through this program. I felt like
[00:22:26] I was like miles ahead of it. But
[00:22:28] sometimes nobody’s going to understand
[00:22:29] that unless you actually have the
[00:22:31] credentials by behind you. And I know um
[00:22:33] we’re going to touch on that a bit a bit
[00:22:35] more and like what are some of the
[00:22:36] challenges especially coming from a
[00:22:39] non-traditional background. There’s a
[00:22:40] lot of pressure around like are you
[00:22:42] ready for this? Do you know what you’re
[00:22:44] doing? Can you handle what I’m going to
[00:22:45] give you? And there’s some skepticism
[00:22:47] from some of the like the old guard of
[00:22:49] people who they went the traditional
[00:22:51] path. They’re used to things being a
[00:22:52] certain way, but also they can get stuck
[00:22:54] in the comfort zone. They can get stuck
[00:22:56] in things must be this way and then they
[00:22:58] don’t modernize and then we end up with
[00:23:00] likenet three systems that Gartner
[00:23:03] predicts 40% of Agentic AI projects will
[00:23:05] fail by 2027 because legacy systems
[00:23:07] can’t support them. What does the system
[00:23:09] actually need before Agentic AI is even
[00:23:11] on the table? as a lot of our
[00:23:13] conversation we’ve kind of talked about
[00:23:15] people and I think there needs to be a
[00:23:18] removal maybe a a re a new approach to
[00:23:21] how companies operate departments make
[00:23:23] sense specialization makes sense but
[00:23:25] when agentic AI is successful it has it
[00:23:28] kind of needs to be multiple people at
[00:23:30] once you have an a so let’s say you want
[00:23:32] to do a task and then as if you were
[00:23:33] hiring a team that AI agent is going to
[00:23:35] be that team and so whoever is building
[00:23:37] it is going to have to go across
[00:23:39] departments across roles across
[00:23:41] specializations to understand things. So
[00:23:43] you can’t have one engineer who knows
[00:23:45] nothing about human resources
[00:23:46] modernizing across a human resource
[00:23:49] role, a financial role, a data role, a
[00:23:51] sales role. We have to have a much more
[00:23:53] truly collaborative environment. And I
[00:23:55] know for many companies from small to
[00:23:57] large, it’s easy to work in silos and
[00:23:59] sandboxes and also to maybe not
[00:24:02] understand each other’s language or how
[00:24:04] each other’s workflow and team processes
[00:24:06] go. So for an agentic AI to be
[00:24:09] successful, it has to have the right
[00:24:10] context across departments, access to
[00:24:12] the right data, the best prompts that
[00:24:14] are going to make it actually work
[00:24:15] effectively, that it’s actually going to
[00:24:17] do the job and not waste a bunch of
[00:24:19] money running in like useless loops and
[00:24:21] spitting out junk, which I think AI can
[00:24:23] be criticized for a lot.
[00:24:25] >> That over the next few years we might I
[00:24:27] think you know saying that most AI
[00:24:28] projects will fail by 2027. Yes. But I
[00:24:31] think in failure we build opportunity.
[00:24:33] What is that saying? You know, you fail
[00:24:35] forward. So, okay, maybe I tried 10
[00:24:37] things, eight of them failed, but the
[00:24:39] two that worked, we now have a model for
[00:24:41] how to do everything else moving
[00:24:43] forward. So, maybe from 2027 to 2030, we
[00:24:46] might actually see a lot more success
[00:24:47] and we’ll probably see companies start
[00:24:50] changing how some of their departments
[00:24:52] work where there will be we hear a lot
[00:24:54] about crossunctional teams, but I don’t
[00:24:56] know if it’s just like a myth. I don’t
[00:24:58] know how well like from the different
[00:25:00] people I’ve talked to cross functional
[00:25:02] teams can be challenging but I hope that
[00:25:04] actually AI not only will improve it but
[00:25:06] that’s what’s going to allow it to work
[00:25:08] and I truly hope that work becomes
[00:25:10] better for people that we’ll enjoy our
[00:25:12] work more because AI is going to do the
[00:25:14] boring stuff and we get to actually
[00:25:15] focus on building team the fun part of
[00:25:17] work so 60 to 80% of corporate data is
[00:25:21] already redundant or obsolete that we’ve
[00:25:23] also talked about at length and most
[00:25:25] organization never hire the people who
[00:25:27] understand the old system alongside the
[00:25:29] people building the new one. Where does
[00:25:32] governance even begin in this criteria?
[00:25:35] >> So, and it also depends, I guess, on the
[00:25:37] industry because different industries
[00:25:38] will have different governance
[00:25:40] standards. But I think we need to get
[00:25:42] sometimes like a middleman. Somebody who
[00:25:44] can communicate between the
[00:25:46] modernization people and then the legacy
[00:25:48] people and bring them together. Someone
[00:25:50] more non-traditional maybe like myself
[00:25:52] even who can help work across the teams
[00:25:54] to make sure that the the old systems
[00:25:57] actually integrate with the new ones.
[00:25:58] Like I think in one of the previous
[00:26:00] questions we went through was finding
[00:26:02] which data belongs on the new systems
[00:26:04] and which data can go on the old systems
[00:26:06] or in some companies you don’t even need
[00:26:08] all of this data like if it’s not
[00:26:10] required by law for you to keep 7 years
[00:26:12] to 10 years of records. It might be
[00:26:14] worth having each team have a directive
[00:26:17] to go through reorganize their drives.
[00:26:19] archive old files or delete archives of
[00:26:22] things where you have copy of copy of
[00:26:24] copy number five and because a lot of a
[00:26:26] lot of companies do a lot of people do
[00:26:28] and so I just think some of it it’s
[00:26:30] funny how often I’ve said the word human
[00:26:32] in a in a question in a discussion about
[00:26:34] AI and about technology but at the end
[00:26:37] of the day all of this technology um
[00:26:38] there are humans behind it and so
[00:26:40] there’s a really famous quote by IBM
[00:26:42] that’s a computer cannot be trusted to
[00:26:44] make a management decision because it
[00:26:46] cannot be held responsible and so at the
[00:26:48] end of the Today so many of these
[00:26:50] questions are answered by the humans
[00:26:52] behind it and not necessarily a pure
[00:26:54] technical or logical process but who is
[00:26:57] the person who can build the
[00:26:58] relationship who can communicate with
[00:27:00] the right terminology to all of the
[00:27:02] teams to get the job done. Well Lydia
[00:27:04] you know you you also mentioned
[00:27:06] previously that probably you have like a
[00:27:08] 10 years age gap from probably most of
[00:27:10] the people in your org at the moment. So
[00:27:12] you know does that reflect at a
[00:27:15] professional level as well where you
[00:27:17] would find that probably they do not
[00:27:19] have they are not as up to date with the
[00:27:22] knowledge in comparison to as probably
[00:27:23] you are or maybe most of the knowledge
[00:27:25] trickles down from probably the CEO only
[00:27:28] and only then people start talking or
[00:27:31] thinking about it. I love this question
[00:27:33] because it goes both ways. I you know
[00:27:35] there are things that I know and then
[00:27:38] trying to communicate it in a way that
[00:27:40] they are going to understand and vice
[00:27:42] versa has been a challenge especially
[00:27:44] when there’s you know 20 years of
[00:27:47] working relationships and a way that
[00:27:49] people communicate a way that people
[00:27:50] operate and I’m coming in and
[00:27:52] introducing a lot of change to that it
[00:27:54] can be disrupting it can be off-putting
[00:27:56] and so I’ve had to one not try not to
[00:27:59] get too excited about all the great
[00:28:01] things we can do to make things better
[00:28:03] because sometimes people take here
[00:28:05] better whatever that is whether it’s AI
[00:28:07] or automation or modernization or just
[00:28:10] even like cleaning up the database it
[00:28:12] can sound like I’m not good enough and
[00:28:14] so working with the people who are more
[00:28:17] seasoned in the company than I am I
[00:28:19] really want to focus on building
[00:28:21] relationships with them so that they can
[00:28:23] see that I’m part of your team it’s been
[00:28:25] a fine balance of hey I want to learn
[00:28:27] from you but how do I learn from you
[00:28:28] without coming off as like hey I’m this
[00:28:30] kid I’m the new kid in the company and
[00:28:32] so there’s a balance between learn from
[00:28:35] me and what I can show you and what I’ve
[00:28:37] learned from you know all of the new
[00:28:38] things that are coming out in the
[00:28:40] universities and you know in the what
[00:28:42] the latest in technology that I can
[00:28:44] bring to the table versus what I can
[00:28:46] learn from them and how did they build
[00:28:47] the system what were the reasons behind
[00:28:49] the way it was built because some of
[00:28:51] that context is also missing in
[00:28:52] modernization when you have a new team
[00:28:54] or a new technology come in they just
[00:28:56] want to update everything without
[00:28:58] realizing there was a reason that things
[00:29:00] and and a logic to how it was already
[00:29:02] made that we don’t want disturb too
[00:29:04] quickly. I’m sure it it makes sense and
[00:29:06] you can probably say it in different
[00:29:07] ways.
[00:29:08] >> Financial services is one of the most
[00:29:09] regulated industries in the world and
[00:29:11] yet AI is being deployed faster than the
[00:29:13] rules governing it. What risks are
[00:29:16] quietly building up in that window?
[00:29:17] >> I’m excited about this question. I
[00:29:19] really love this question because not
[00:29:21] only in the financial services industry
[00:29:23] but also in government, healthcare and I
[00:29:25] think a lot of just other tech companies
[00:29:28] or public facing companies some there’s
[00:29:30] a gap. So some people are using AI way
[00:29:33] too fast. Just like you have, you know,
[00:29:35] Debbie often social services who’s
[00:29:37] uploading case files into chat GBT to
[00:29:40] help her do her work. That’s a huge
[00:29:41] risk. And some people just don’t see it
[00:29:44] yet. They don’t see that you cannot
[00:29:46] upload your, you know, personal
[00:29:48] information into this LLM, which is
[00:29:50] essentially being stored in a database
[00:29:52] somewhere that can be hacked. That is
[00:29:53] maybe is you’re sharing personal private
[00:29:55] information with an external third party
[00:29:58] that is technically against regulations.
[00:30:00] But in financial services, in my
[00:30:02] industry specifically, we have a lot of
[00:30:05] rules about personal information, how it
[00:30:07] should be shared, how it should be
[00:30:08] accessed or, you know, really not shared
[00:30:10] at all. It can there there’s a lot of
[00:30:12] information that I deal with and my team
[00:30:14] and everyone in our company deals with
[00:30:16] every single day that is extremely
[00:30:18] personal and private. It’s not as much
[00:30:20] as I guess healthcare would be
[00:30:22] considered, but there’s a saying that’s,
[00:30:24] you know, you don’t mess with people’s
[00:30:25] kids and you don’t mess with their
[00:30:26] money. So, one of the things that we’ve
[00:30:29] done or that I took a step towards doing
[00:30:31] when I first saw AI coming out, I felt
[00:30:33] like I was very ahead of the curve on
[00:30:34] this because nobody understood what I
[00:30:36] was talking about when I kept telling
[00:30:38] about the risks of using AI with
[00:30:41] anything especially in our industry
[00:30:42] because that personal information. So
[00:30:45] what we did and we created a platform
[00:30:47] specifically for our company and for our
[00:30:49] office users that you can connect to
[00:30:50] claude chatpt any LLM but we put a
[00:30:54] tokenization layer into the prompt
[00:30:56] window that essentially identifies it’s
[00:30:59] based off of Microsoft’s Presidio
[00:31:00] technology and it identifies any PII
[00:31:03] such as even anything from and you can
[00:31:05] choose like the the strength level but
[00:31:07] it can be very high or very low but it
[00:31:09] will recognize credit card social
[00:31:10] security number name email data that you
[00:31:12] don’t necessarily want to go into a
[00:31:15] third party company and then tokenize it
[00:31:17] in transit and then bring it back
[00:31:18] decrypted. And so the user doesn’t see
[00:31:21] this process. They don’t feel the
[00:31:22] process. But it essentially prevents our
[00:31:25] u clients or our agents information from
[00:31:28] going to external sources and being used
[00:31:30] to train other models because I think in
[00:31:33] April of last year or was it time flies
[00:31:35] I think it was April 2023 Open AI was
[00:31:38] hacked. I mean this wasn’t like widely
[00:31:40] advertised but just to show that even
[00:31:42] some of these companies that we think
[00:31:43] are so big and untouchable they do get
[00:31:46] hacked. Data does get leaked and then at
[00:31:48] the end of the day we are liable for how
[00:31:50] we use their platform. So um I was
[00:31:53] really proud of this um development. I
[00:31:56] think now I’ve seen some companies that
[00:31:58] specialize in this type of PII
[00:32:00] identification and protection when using
[00:32:02] LLM, but the legislation will have to
[00:32:05] keep up. like how can we utilize it has
[00:32:07] not been clearly defined. So we try to
[00:32:09] be very careful and we’ll see we’ll see
[00:32:12] how how things develop.
[00:32:13] >> Post migration organizations tend to
[00:32:15] celebrate the finish line too early.
[00:32:17] What’s the metric that actually tells
[00:32:19] you it worked not the one that it
[00:32:21] reports up with but the one that sort of
[00:32:23] really matters.
[00:32:24] >> I think it comes down to users and sales
[00:32:27] specifically for appreciation financial
[00:32:30] we if we have people using it what so
[00:32:32] for us our people are our agents. If we
[00:32:34] introduce something and it’s being used,
[00:32:36] then we know it’s working. We know it’s
[00:32:38] good. If we have increased sales, we
[00:32:40] know it’s good. If we know nobody’s
[00:32:42] using it and sales are dropping, it’s
[00:32:44] not working. So for us, that’s probably
[00:32:46] a very simple straightforward metric,
[00:32:49] but I think it’s a total sign. According
[00:32:51] to Brookings, 86% of the workers are
[00:32:54] most exposed to AI replacement are
[00:32:56] women. And yet, according to IMD, women
[00:32:59] hold fewer than 14% of senior AI
[00:33:02] leadership roles. as someone building
[00:33:04] the technology and sitting in the
[00:33:05] leadership seat herself. How do you sit
[00:33:08] with this ongoing tension? I feel that I
[00:33:10] I think honestly I have seen advocated
[00:33:13] for replacement of certain roles where I
[00:33:15] was a woman in the role and it all came
[00:33:18] down to economics of I would encourage
[00:33:20] women to use AI to make their work
[00:33:22] better, faster, stronger and stand out
[00:33:24] because at the end of the day if your
[00:33:27] job is if you cost $200 an hour and
[00:33:31] you’re producing work that an AI model
[00:33:33] can do in mere seconds as a company
[00:33:36] leader, I cannot justify the cost. They
[00:33:39] cannot sit with the CFO and the CEO and
[00:33:41] review the budget and review the numbers
[00:33:43] and say we’re going to keep her on theft
[00:33:46] just because we like her. At the end of
[00:33:48] the day, companies won’t run like that.
[00:33:50] And it’s really sad because I have seen
[00:33:52] people I love and care for dearly lose
[00:33:54] their jobs um for from these type of
[00:33:56] decisions. And unfortunately, I have
[00:33:57] pushed for some of these decisions where
[00:33:59] I could see I cut a 20 I cut a $20,000
[00:34:03] budget for some parts of our company
[00:34:05] because I said AI can do this. I feel
[00:34:07] personally guilty um when I see some of
[00:34:10] the results of maybe a more stressed out
[00:34:12] department leader who doesn’t feel
[00:34:14] prepared to take on using AI for her
[00:34:16] department now has had a massive budget
[00:34:18] cut and then the the people who think
[00:34:21] that AI should not be replacing content
[00:34:23] creation or copywriting are feeling very
[00:34:26] snubbed and actually becoming even more
[00:34:29] anti- AI. In some ways it has the
[00:34:31] opposite effect of what I intended for.
[00:34:33] Um, so how I’m dealing with it is to try
[00:34:36] and step in to work more closely with
[00:34:38] the people that I’m impacting in those
[00:34:41] decisions to show them ways and
[00:34:43] opportunities that they can increase
[00:34:45] their value with AI so that it doesn’t
[00:34:47] become a question later on of are you
[00:34:49] worth what we’re paying for this service
[00:34:51] because AI can do it faster. But then
[00:34:53] you know I also wanted to maybe um make
[00:34:56] it more precise by asking you this other
[00:34:58] question um which adds to it that do you
[00:35:01] think that this is a him or a her issue
[00:35:03] or a generational issue perhaps you know
[00:35:06] with the tendency of people not being
[00:35:08] able to adapt to the tools or maybe
[00:35:10] self-learn or being able to use AI or
[00:35:14] maybe just rely entirely upon an AI
[00:35:16] model and not have a skill set of their
[00:35:19] own because in my experience perhaps
[00:35:21] I’ve noticed that it is at times a
[00:35:24] generational problem too with Gen Z
[00:35:26] perhaps.
[00:35:27] >> Absolutely. I think this is an excellent
[00:35:29] point where some people look at is it a
[00:35:31] gender issue. I think saying it’s a
[00:35:33] generational issue is much more accurate
[00:35:36] because we had to we did let someone go
[00:35:38] who was you know from much like from an
[00:35:41] older generation I should say but he
[00:35:43] used to be like a rocket scientist. He
[00:35:45] used to work on top secret projects but
[00:35:48] he was not willing to adapt to the new
[00:35:51] systems to the new models and it became
[00:35:54] a liability. And I’ve definitely seen
[00:35:56] the complete opposite where there’s some
[00:35:58] people who are younger and all they do
[00:36:00] is AI and sometimes it’s overproducing
[00:36:03] like you I’m sure you’ve seen when you
[00:36:05] prompt let’s say Claude for something
[00:36:06] and then it produces a 13page report.
[00:36:08] Nobody’s going to read those 13 pages
[00:36:10] cuz they lack a lot of context. it
[00:36:12] produces kind of ill relevant or
[00:36:14] irrelevant parts that we don’t really
[00:36:16] need to see. So there needs to be some
[00:36:18] middle ground. And I think you know I
[00:36:21] think training and and who is going to
[00:36:23] be the trainer is something that I often
[00:36:25] wonder is like I don’t necessarily have
[00:36:27] time to be the trainer. Do I even trust
[00:36:29] a lot of the consultants who say they
[00:36:31] can train my company because they don’t
[00:36:32] know our business? They don’t know what
[00:36:34] we need cuz I’ve seen um some of my best
[00:36:37] team members don’t use AI at all. I’m
[00:36:39] thinking of one person particularly if
[00:36:41] she sees this podcast she’ll know who
[00:36:43] she is. She is so detail oriented like
[00:36:45] she has like laser vision. She can catch
[00:36:47] everything. And you still need people
[00:36:49] like that to catch the AI issues who are
[00:36:52] willing to look at the details because I
[00:36:54] have lots of people in the younger
[00:36:57] portion I guess of our team who they’ll
[00:36:59] use AI no problem but it’s easy to
[00:37:01] overlook some of the mistakes that it
[00:37:03] makes. So I would say that tension can
[00:37:05] easily be managed by teamwork. I think
[00:37:08] teamwork is one of the most important
[00:37:10] things to have in a company to help
[00:37:12] remove silos, go outside of sandboxes
[00:37:16] and have real discussions across you
[00:37:18] know the generational age gaps and even
[00:37:21] sometimes the not just department but
[00:37:23] like seniority cuz sometimes the senior
[00:37:25] leaders they know everything they have a
[00:37:27] vision but don’t necessarily get the
[00:37:29] feedback from the field or from the
[00:37:31] people that are working dayto-day on the
[00:37:33] ground and to just have more checks at
[00:37:36] every level. You know, great answer.
[00:37:37] Moving towards the last question, Lydia.
[00:37:39] What’s the one thing you’ve had to
[00:37:41] unlearn on this journey that you wish
[00:37:43] someone had told you earlier? There’s a
[00:37:45] few things. There’s a lot of things I’ve
[00:37:46] learned and then what have I unlearned?
[00:37:48] And some of it is that the the way that
[00:37:51] we used to work of like you need to be
[00:37:53] an expert on this one thing and like
[00:37:55] that’s it. Not how things are.
[00:37:56] especially at the leadership level as
[00:37:58] I’ve been and gratefully so been put
[00:38:00] into more and more of a leadership from
[00:38:02] my curiosity to learn things is that
[00:38:05] things should not I really don’t think
[00:38:06] things should be sandbox departmentally
[00:38:08] there needs to be more openness there
[00:38:10] needs to be more cross collaboration in
[00:38:12] some ways I’ve also had to step back
[00:38:15] from some of my my openness personally
[00:38:18] to be a leader I’ve also had to not be
[00:38:21] so accessible not be so available to
[00:38:24] help to set very firm boundaries with
[00:38:26] people sometimes as well. And one of the
[00:38:28] things that makes me think about that is
[00:38:30] when you use the term resource because
[00:38:31] some people at the top level of a
[00:38:33] company, everything is a resource.
[00:38:35] Whether you’re a human or whether you’re
[00:38:37] the talking about cloud storage, all
[00:38:39] these are resources at the end of the
[00:38:41] day that help the company grow and
[00:38:42] operate. And so when you remember that
[00:38:44] the people that you’re dealing with
[00:38:45] sometimes they just see you as a
[00:38:48] resource and if you’re not meeting their
[00:38:50] needs, then you are no longer a resource
[00:38:52] that they want to use. And it’s not
[00:38:54] human. And I’ve always focused on
[00:38:56] especially with my work at the people’s
[00:38:58] accords which we didn’t talk about on
[00:38:59] this but anyone who’s watching this is
[00:39:02] welcome to go look up the people’s
[00:39:03] accords and learn more about how we use
[00:39:05] technology and more of a peoplebased
[00:39:08] startup startup type of mission so we
[00:39:10] can maybe even meet again to talk about
[00:39:12] that. I would love to. but to take away
[00:39:14] some of that human element and just be
[00:39:16] more willing to be black and white and
[00:39:18] to balance that when to know based on
[00:39:20] the situation of should I call the whole
[00:39:22] team in for a meeting and work on this
[00:39:23] together or do I need to say nope we’re
[00:39:26] only having two people on this call and
[00:39:27] we’re going to get it done it’s it’s I
[00:39:29] know the question was about unlearning
[00:39:31] but I find it to be more of adapting how
[00:39:34] have I learned to adapt to situations
[00:39:36] more quickly instead of having like a
[00:39:38] solid playbook
[00:39:40] >> no I think that that was absolutely
[00:39:41] fascinating thank Thank you, Lydia, so
[00:39:43] much for being super honest with us and
[00:39:45] our listeners. This was a great
[00:39:46] conversation.
[00:39:47] >> Thank you, Rabia, for having me. I
[00:39:49] really appreciate it and um I look
[00:39:51] forward to talking to you in the future.
[00:39:54] [music]
