3121: Gen AI at Work: BCG and OpenAI's Study on Expanding Employee Capabilities

[00:00:03] Welcome back to the Tech Talks Daily Podcast. In today's episode, we're going to talk about the transformative potential of generative AI and how it can reshape how knowledge workers approach very complex tasks.

[00:00:17] Because recent research from the BCG Henderson Institute conducted in collaboration with Boston University and OpenAI's Economic Impacts Team has shed light on how tools like ChatGPT are already impacting things like productivity, skill development and workflows.

[00:00:38] And this comes at a time where there's been a lot of question marks around the ROI of AI and also what tangible or measurable difference AI can have on the workplace.

[00:00:50] Some businesses have gone headfirst thinking about the technology rather than the problems they want to solve.

[00:00:56] And that is more to do with why they're not getting the results that they require, in my humble opinion.

[00:01:02] But I think from with a more focused approach in anything from enabling non-experts to tackle coding or predictive modeling to enhancing the capabilities of seasoned professionals.

[00:01:14] The findings in this report present a fascinating glimpse into a future of AI powered workplaces.

[00:01:21] So today I invite you to join me as we try to unpack some of the insights from this report and discuss what they mean for the future of work, talent management, executive strategies, education, so much more.

[00:01:35] We've got a lot to get through today.

[00:01:37] So enough scene setting for me.

[00:01:39] It's time for me to introduce you to today's guest.

[00:01:43] So a massive warm welcome to the show, Vlad.

[00:01:46] Can you tell everyone listening a little about who you are and what you do?

[00:01:50] My name is Vlad Lukic and I run tech and digital practice at BCG.

[00:01:54] So my team advise this CEOs that want to figure out what should be the technologies they focus on and how to build their roadmaps to deploy them most effectively in their businesses.

[00:02:06] We also advise CIOs, CTOs, chief digital officers as they are articulating their roadmaps, modernizing their technology, deploying the latest technology like Gen.AI in their operations.

[00:02:18] And then we also help various other executives that are trying to deploy tech in their daily lives and in their daily operations.

[00:02:24] So it's been a fascinating journey the last few years.

[00:02:28] I bet it has.

[00:02:29] I mean, we're talking about what technology businesses should be investing in, should be adopting and implementing.

[00:02:36] Predictably, I would imagine over the last two years, there's going to be a lot of conversations around AI.

[00:02:42] And I've got to ask, I mean, what would you say are the most surprising findings from BCG's experiment on Gen.AI's impact on consultants' ability to perform data science tasks?

[00:02:54] Because I was reading about this recently, and I'd love to hear more about the actual results, because another big topic right now is ROI of any technology.

[00:03:02] So what did you find here?

[00:03:04] Yeah, so it's interesting.

[00:03:05] And a little bit of the backdrop.

[00:03:07] We've taken a control group of the data scientists, then a group of consultants without any background in data science, and then a group of consultants without any background in data science, and given them the Gen.AI tools to conduct some specific tasks in code writing, doing some predictive analytics, and then doing statistical analysis.

[00:03:31] And then we tried to compare the results.

[00:03:35] And what was fascinating was that for those who had no background in data science, they improved their performance dramatically out of the gate by using these tools.

[00:03:48] What was more surprising was that those who were experienced when given these tools, they performed even better, because we just thought they're already performing at an excellent level.

[00:04:00] They actually performed even better when these tools were given into their hands.

[00:04:04] And then the third thing that was, I would say, surprising was think of it as an exoskeleton.

[00:04:10] When we were done with the study or done using the tools, if we asked the non-experienced folks to not perform those tasks without using the tools, they couldn't, because it took the exoskeleton away from them.

[00:04:22] So they didn't learn how to do the work.

[00:04:25] They just had something that allowed them to do it.

[00:04:27] So it was just, those are, for example, three things that stood out as we did the work.

[00:04:33] Wow, that's amazing.

[00:04:34] So if we were to take those individuals that had no prior coding experience at all, how did Gen.AI enable them to complete complex tasks like, I don't know, predictive modeling or statistical analysis?

[00:04:49] And if you can share around what it enabled them to do and how it did that?

[00:04:54] Yeah.

[00:04:55] So let's say, for example, on code writing, right?

[00:04:57] If you did not do the code writing before, if I asked you not to write a piece of code, you would have to probably go online.

[00:05:04] You would go on YouTube.

[00:05:05] You would look at some of the instructions of how you want to go about writing the code here.

[00:05:09] They could just prompt what they're looking for, right?

[00:05:11] And they could get the first wave of the code.

[00:05:14] They could then follow up with questions on how do I deploy it?

[00:05:17] How do I make it executable, et cetera?

[00:05:18] So in a conversational way, they suddenly became able to develop code.

[00:05:25] Now, what we're also finding, not necessarily in this study, but as it gets extrapolated and we test it in real life, this works really well for straightforward, plain vanilla, down the fairway, pick your favorite expression tasks, where if you are non-coder, you can ask questions and you get code that is pretty well written and can be used.

[00:05:46] When you get to more complex and exotic tasks that require a bit of fine tuning of the code, that's where it becomes tricky for a novice because they will ask for a piece of code that comes in, let's say they get 30, 40 lines of code.

[00:06:01] It's hard for them to even sanity check.

[00:06:02] Is it good or not?

[00:06:03] This is where more experienced coders, they know how to ask the question.

[00:06:07] So they chunk it up.

[00:06:08] They say never more than three or four lines of codes.

[00:06:10] They can chunk up the prompts in a way that is easy to sanity check and then they can string it together, right?

[00:06:18] And so this is where it becomes, when you combine this study with other observations, some more broader, interesting learning start popping up.

[00:06:27] So the standout thing there really seems to be that it enables anybody to become able and have those skills.

[00:06:34] But when it comes to refining or retaining skills, it gets a little bit trickier.

[00:06:39] So what would you say the limitations are of using Gen AI for reskilling?

[00:06:44] And how can organizations better balance those immediate capability gains that we're talking about here with those long-term learning and skill retention that is also desirable on a longer-term basis?

[00:06:57] Yeah.

[00:06:58] Um, but there's, before I answer that question, there is, it prompted a thought on one other observation that was very interesting.

[00:07:04] Those who had coding experience prior to this type of work, uh, their base capabilities and, and quality at which they were solving all the problems was 20, 30% higher than those who didn't have coding background.

[00:07:19] Even if they were asked to do tasks that are unrelated to coding, which was very interesting.

[00:07:22] So to your question, there is one piece, which is, um, as you're hiring people for these, um, in this new world, right, where we need constant upskilling and learning, looking for people with coding background actually matters quite a bit.

[00:07:36] Cause they actually have in their skillset ability to try new things and base load of knowledge on top of which they can build, um, to use some of these new tools.

[00:07:45] So that's kind of one, uh, one interesting observation, um, in terms of, um, like, in addition to like getting the right people in, uh, once you have them in the, the limitations of the Gen AI is, um, don't automate all of it.

[00:08:00] Right.

[00:08:00] It really should be viewed as a co-pilot as an ad, as a support.

[00:08:06] I mentioned word exoskeleton as something that makes you actually be more effective in the tasks that you're doing.

[00:08:13] Right.

[00:08:13] Um, and as you're thinking about upskilling the organization, um, you just need to be aware of the limitations and that it is a tool that can aid.

[00:08:22] Right.

[00:08:22] Right.

[00:08:22] And therefore, um, a be aware of the limitations to understand the whole workflow.

[00:08:29] Right.

[00:08:29] And understand where that aid can be helpful and don't automate those steps.

[00:08:35] Right.

[00:08:36] Um, so you can actually use it to speed up the idea generation.

[00:08:40] You can use it to speed up the first drafts of the documents of the code, et cetera, but then you got to design in the process, experience people that can quickly sanity check.

[00:08:50] Right.

[00:08:51] So that's kind of, I would say second observation.

[00:08:53] The third one is, um, anthropologists in me is like loving it for the last nine months.

[00:08:59] Cause there's a lot of these little lessons on how do you manage your teams and how do people interact?

[00:09:04] Right.

[00:09:04] So the third lesson there is, um, have the experienced folks in the workflows and we call it force purposeful toil into the steps.

[00:09:15] Right.

[00:09:16] Right.

[00:09:16] So one of the taglines we use is Gen.ai is great to eliminate, uh, toil out of the work, right?

[00:09:22] The things that we don't want to do, you can use Gen.ai for, but we're learning, you got to force purposeful toil in there where you force some steps and reviews for the less experienced folks.

[00:09:33] So they learn how to sanity check things and they learn, and you help them build intuition, um, how to even go about answering some of the questions.

[00:09:42] So, uh, for that means like in this example of encoding, don't allow the juniors to do more than six lines of code using Gen.ai.

[00:09:51] Don't just prompt.

[00:09:52] And then you get 50 lines of code that you cannot unravel or sanity check.

[00:09:56] So you force them, you force them then to when they have 30 lines of code to engage with their manager, right.

[00:10:02] To see how it comes together.

[00:10:03] So you can force some of the interventions inside so that you can retain some of the knowledge and build some of the intuition for those non-experienced folks.

[00:10:11] Does that resonate?

[00:10:12] Does that make sense?

[00:10:14] Yeah, completely.

[00:10:14] And I think you use the right words that co-pilot collaborate support.

[00:10:19] And I think very often we get, it's, we see a lot of news articles about, um, machine versus human.

[00:10:27] And I think it's all about getting or working together to get a little bit further along the road in ways that neither machine or human is able to do on their own.

[00:10:37] But when they work together, they can get so much further and so much faster as well.

[00:10:42] And how do you see Gen.ai's role as almost like a brainstorming partner that enhances problem solving and collaboration for tasks like predictive analytics or, or indeed any area in the workplace?

[00:10:55] Because again, that, that seems to be something that's being leveraged for more and more.

[00:11:00] Yeah, it's fantastic for those, those types of tasks, right?

[00:11:03] To give me a first wave of ideas for the following thing.

[00:11:06] Right.

[00:11:06] And I did a mini experiment with my own teams, um, where let's say we would get like 50 or more interviews on a specific topic.

[00:11:14] And then I would have my individuals, um, give me the summary of them, right.

[00:11:18] Or key themes, give me top two, three themes.

[00:11:20] Right.

[00:11:20] And I would take it from five different individuals.

[00:11:24] They would all on average get the top two themes, but then the third four theme, each of them would pick up on a different one.

[00:11:31] Right.

[00:11:31] And then I would use Gen.ai tool and say, Hey, give me the summary of the themes, which didn't happen literally in a second.

[00:11:37] And it would usually give me the most comprehensive list.

[00:11:41] So I would then get five things that are actually all there, right.

[00:11:45] That were mentioned by all the individuals, but each individual got only two of them there.

[00:11:50] Uh, the tool actually picked up on most of those themes.

[00:11:53] Right.

[00:11:53] And then I went and did my own research or reread through the materials.

[00:11:57] I actually picked up on one or two things that are very nuanced things that none of them could pick it up.

[00:12:02] They were not spoken.

[00:12:03] It was actually combo of the things.

[00:12:05] So this is where, um, like to, to build off of the previous, um, examples, the Gen.ai tool gave me the most comprehensive list up front to get started with.

[00:12:15] But both them and inexperienced folks didn't pick up on the nuances and connective tissue between the things.

[00:12:21] So there's still work for the experienced folks to then connect the dots and add item six and seven into the, uh, into the lists.

[00:12:29] Right.

[00:12:29] And so this is where the combo needs to be in place.

[00:12:32] Yeah.

[00:12:33] Well, I think I was reading on Reddit a few weeks ago.

[00:12:35] There's a telegram group out there where people are using it to turn their ideas into businesses.

[00:12:40] And there was a number of them there that were using Gen.ai, uh, to generate their business idea or expand on their business idea.

[00:12:49] And then let's say it was a t-shirt company and then it was what kind of t-shirts.

[00:12:53] It told them what t-shirts to make, how to get them made, where to get them made, how to build the website, how to market the idea.

[00:13:01] It just brings to life that, that almost co-pilot, uh, theme that we're talking about here.

[00:13:06] And you can bring anything to life if you bounce ideas and use it to brainstorm.

[00:13:11] Right.

[00:13:12] Yeah, it's, it's, um, so I have three kids.

[00:13:15] Uh, and so you can imagine conversations can get, they're all under 18, so you can imagine conversations and can get wide ranging, but it became such a wonderful co-pilot for us.

[00:13:25] Um, I call it the sixth member of the family when the kids are like, oh, look at that ice cream stand.

[00:13:30] Like, I wonder, like, like, would that be an interesting business to run?

[00:13:34] And like, you won't even make money.

[00:13:35] And then we were like, okay, so if I'm to run an ice cream stand in Boston, in Boston common, what should I expect of it to, as a business book in terms of margin that I could make and, and overall dynamics enter.

[00:13:48] And then it gives me like, listen, given you're in Boston, you probably have only three months in the summer.

[00:13:52] If you're in Boston common, that means, bah, bah, bah.

[00:13:55] If you have one stand, you're probably limited to this amount of ice cream on a daily basis.

[00:14:00] Therefore, dah, dah, dah, dah, dah, if you price it on these levels, dah, dah, dah, dah, dah, suddenly tells me, hey, the, the margin you can expect from a, from a stand like this is $500 on a daily basis over this period.

[00:14:12] You also need to keep in mind your capex investments.

[00:14:14] It gave me a whole business case upfront.

[00:14:17] My kids looked at it like, this is a hard business to run.

[00:14:20] Um, uh, on the flip side, the other one said, well, what if I have a hundred of these?

[00:14:24] Well, suddenly it's an interesting business to run.

[00:14:26] But then, so it became an aid to have that conversation quickly.

[00:14:30] With the kids on a pretty complex topics, right?

[00:14:32] On like running a business.

[00:14:33] Right.

[00:14:34] Um, and so just to bring to life, Neil, your point, like it is just such a wonderful, um, aid, right.

[00:14:41] And a thought partner in a lot of these things.

[00:14:43] Right.

[00:14:43] And if you, if you have a bit of intuition and know how to sanity check it, it just, it helps you speed up a lot of those things.

[00:14:50] A hundred percent with you.

[00:14:52] And I love that.

[00:14:52] I mean, just getting your kids to think bigger and think about how to run a business and different ideas just from observing something in the street there.

[00:15:00] And I've, I've seen so many other ideas as well.

[00:15:03] Planning trips, family holidays.

[00:15:05] You can literally say, I want to go to this location.

[00:15:08] Where should I stay?

[00:15:09] What sightseeing trip should I do?

[00:15:11] What are the best restaurants?

[00:15:13] And then export it straight to a PowerPoint presentation to show your, your family that the opportunities are endless around that.

[00:15:19] So anybody that is sat on the fence, I do encourage them just to go out there and have a play with it.

[00:15:24] And, uh, it's the best way of getting to grips with it, but I mean, bring it back into the workplace.

[00:15:29] What, what implications do your experiments results have for refining expertise and talent management strategies within their organizations?

[00:15:39] Anything you can share around what you found there?

[00:15:42] Well, listen, it's, it's, uh, helping companies, um, start rethinking.

[00:15:48] What does it, you know, what, what is the entry level skills that I need?

[00:15:51] What is the minimum that I need?

[00:15:53] That gave the example of coders already performing at a different level, right?

[00:15:57] Ooh, okay.

[00:15:58] Maybe those are profiles we need to get in, right?

[00:15:59] Because they will be nimbler to do these things too.

[00:16:02] It is asking a lot of companies.

[00:16:04] Okay.

[00:16:05] If I have thousand open positions for new folks, do I need thousand or should we start first with hundred and have them use these tools to get started with?

[00:16:12] And then, uh, we can think about it differently.

[00:16:15] It's also interesting to me.

[00:16:17] There was a lot of fear up front in the early days.

[00:16:22] Hey, will AI take over the humans?

[00:16:25] Right.

[00:16:26] And there will be no jobs, et cetera.

[00:16:27] Job growth is up.

[00:16:29] Right.

[00:16:29] And when I engage with our clients, the earliest adopters and those that are actually deploying it with best results are all with the growth mindset.

[00:16:36] Like I cannot physically keep growing at the pace we would like to, uh, if I have to add human for all the role that we have.

[00:16:44] I really need this exoskeleton or something that supercharges every single new employee that I'm bringing in.

[00:16:50] Right.

[00:16:51] So even the training is going in that direction of how to infuse this into the workflows, both in onboarding, but also on an ongoing basis.

[00:16:58] It has impact on how you manage the workforce, how you onboard the workforce, how you think about upskilling, et cetera.

[00:17:06] And for anybody listening that is concerned around that ROI of AI or how does it work in my workplace?

[00:17:14] How can I use it to make a tangible and measurable difference in my workplace and refine processes, et cetera?

[00:17:20] Any advice you could offer on how businesses can effectively integrate Gen.AI into their, their current workflows and maybe even expand employees' technical capabilities while ensuring sustainable skill development?

[00:17:34] I appreciate that question on its own is almost an entire podcast episode, but just to get people thinking differently, any advice around that?

[00:17:42] Yeah.

[00:17:42] So listen, first I would say make it available, right?

[00:17:44] In a safe, responsible way with proper guardrails, et cetera, but make it available to your employees to use it, right?

[00:17:50] Cause it's there.

[00:17:51] And what I'm finding out is if you're not making it available, they use it themselves.

[00:17:56] Then they just use it in any responsible way on their private devices, giving your data out, et cetera.

[00:18:02] So you cannot avoid that.

[00:18:03] So just make sure that it's accessible.

[00:18:04] They can have access to these tools.

[00:18:07] They can have access to, um, do, do know where they end up freeing up the time and then have explicit conversations of how to redeploy that time.

[00:18:17] Right.

[00:18:17] Versus, um, that, um, they just be, fill it up with extra other busy work.

[00:18:23] Right.

[00:18:24] And so great examples for that for me are, um, there are some tasks where people do them now in a day versus the whole week.

[00:18:33] Right.

[00:18:33] They finish it and then they're waiting because like, there's no task until next week, but they have no one to have that conversation with around what do I use the time for?

[00:18:43] So they just stretch it around.

[00:18:45] And the, which gets me to the third point companies need to think about a giving the tools, having people start using them and then having explicit conversations on how do we change the whole workflow.

[00:18:55] Right.

[00:18:55] And so we have an example where we took one piece of analysis in, and from 10 days of work, we brought it down to two minutes.

[00:19:05] Right.

[00:19:05] Like we literally from, you know, 10 days down to two minutes, you, you fine tune it and you're done.

[00:19:11] But the customer gets output of that still only after 10 or 11 days.

[00:19:16] Right.

[00:19:16] So we just talk with the team and say, what is going on?

[00:19:19] Like, is the tool not good?

[00:19:20] Is the output not good?

[00:19:21] They're like, no, it's actually very good.

[00:19:23] Like the quality of this is like acceptance rate is like 90, 95% with what we give to the committee to review.

[00:19:30] And we're like, well, why does it still take 10, 11 days for people to use it?

[00:19:34] And the answer is, well, the committee meets only once every 10 days.

[00:19:37] So we cannot even review this earlier.

[00:19:41] So, okay.

[00:19:41] So, but why is the committee even in the place or because the quality that was coming in before was, we accepted only 20, 30% of it when humans did it.

[00:19:49] It was really poor quality.

[00:19:50] So they had to review it and send it back in the process.

[00:19:53] And so, so the quality is now 90% plus it's done in minutes.

[00:19:58] And do we need the committee?

[00:19:59] Well, that's a good question.

[00:20:01] We're not, we probably don't.

[00:20:03] Great.

[00:20:03] But it takes six months to get to that conversation.

[00:20:05] Right.

[00:20:06] And so how do you speed up those types of conversations around how we do the work and how do we hand over tasks to each other?

[00:20:13] How do we redeploy that time?

[00:20:15] Right.

[00:20:15] And let's move the bottlenecks.

[00:20:17] Right.

[00:20:18] And shift them.

[00:20:19] If we have way to do things faster, muscle of doing that, Neil, just kind of evaporated over time.

[00:20:24] Because we, as companies, as a kind of, again, anthropologist in the, in the inner anthropologist in me, we got into this mode of one to 2% continuous improvement mindset.

[00:20:36] We're talking now about 30, 40, 50% improvement.

[00:20:39] And just the muscle of digesting and doing it is, is atrophied over time.

[00:20:43] So that's where I would steer the companies to build that muscle to run that operating type of constant evolution.

[00:20:50] That is, that is exciting.

[00:20:53] I love that.

[00:20:54] And also love the example you gave a few moments ago, just going back to that way.

[00:20:58] You were talking about your kids and the positive impact that Gen AI was having on them and how it sparked that family discussion after seeing that ice cream stand there.

[00:21:09] And so, and on the, and on the trip that you mentioned, we actually did a trip around the world this summer, all five of us and for like nine weeks.

[00:21:16] And this was a companion.

[00:21:18] This is part of the trip, right?

[00:21:20] So both in planning the stops, even when we had guides in some places, I would sanity check what they're asking and would add new data facts into the mix.

[00:21:28] And I'm like, Ooh, okay.

[00:21:29] It elevated the conversation, what we learned to a whole new level versus just walking through and doing the fact.

[00:21:34] So it just to amplify your point from earlier, it's such a wonderful, uh, uh, add on, right.

[00:21:41] To take the experiences to a whole new level.

[00:21:43] And just listening to that story there and the impact it's had on, on not just your children, but you as a family and how you work together and obtain information and enjoy a perfect family holiday.

[00:21:54] It's only a great thing, but in certain elements of education, we keep seeing the, the horror stories of, of, of, of, it's gotta be completely banned.

[00:22:02] And, uh, and I know there's a few reasons behind that, but what lessons do you think from your own experiences and indeed the study, uh, means for educators or parents about the importance of fostering and engineering a mindset, not just through coding and problem solving skills, but how they use, um, gen AI responsibly as well.

[00:22:22] When you put it all together, any advice on that?

[00:22:25] Because we will have a lot of educators and indeed concerned parents listening.

[00:22:29] Listen, I would say you cannot avoid it.

[00:22:31] You need to engage with it and learn both the limitations of it and what it can do.

[00:22:35] And then also you need to not use it in some areas where you need to build intuition, right?

[00:22:40] For some people, for folks to do the, the raw, um, thinking skills.

[00:22:44] So like you triggered, uh, a funny anecdote with that question, which is I was at a dinner with, with educators, uh, with few English teachers.

[00:22:52] Um, and just listening to them, one was saying, we have to ban all this cause this is silly.

[00:22:58] People, kids write these essays on these things and I can see that it was written by someone else or by the machine.

[00:23:04] Right.

[00:23:04] Uh, so we just got to ban it.

[00:23:06] The second one was, you know what?

[00:23:07] We cannot ban it, but I'm using this to, um, I'm building my workflows to assess things that come in.

[00:23:13] So I'm using it to now assess who did it by hand versus who actually used the tool, et cetera.

[00:23:20] So it's okay.

[00:23:21] Let them use it, but I will, I will use the tools to prevent them.

[00:23:24] Right.

[00:23:24] The third one was saying, well, they're going to use it anyway.

[00:23:27] So I'm gonna, I'm telling them use it, but then write four essays and then use it to, and then please analyze those essays and write to me what's the difference between them so that I force them to think about it.

[00:23:39] And it kept going on.

[00:23:40] Everybody had a variation on a theme and then it came to the last one.

[00:23:42] I have no idea what you're all talking about.

[00:23:44] Like, um, I just give them a piece of paper and a pen and say, right.

[00:23:49] Yes.

[00:23:50] Right.

[00:23:50] It's like, oh, okay.

[00:23:52] So, uh, like there is so many different ways to think about it.

[00:23:56] Right.

[00:23:56] Um, and like, if I need you to learn some tasks, right, I can mechanically just give a piece of paper and a pencil and, or I can replicate that in a conversation.

[00:24:05] If I, if I have the technology there and you're going to be using it, then I better be playful with it and have you learn about it.

[00:24:11] So we can jointly decide what the limitations are and how do you interpret the outcomes, how to build the intuition around it.

[00:24:18] So like, uh, key is like be playful with it, use it, no delimitations, engage in the conversation around it.

[00:24:27] And then we'll jointly kind of fine tune what's the best way to deploy it.

[00:24:31] Fantastic advice.

[00:24:32] And you perfectly highlighted there, how there's a team of educators, all with completely different takes, opinions and approaches.

[00:24:40] And I would imagine it's exactly the same in the workplace.

[00:24:43] So any tips on what steps executives should be taking to prepare their organizations for a future where tools like Gen.

[00:24:51] AI will significantly augment and maybe even reshape the workforce, but, uh, inside the workplace, any advice there?

[00:25:00] So, uh, executives themselves should start using it.

[00:25:03] That's one, uh, even if it's small things, they have executives that say, Hey, please monitor our conversation and give us comments.

[00:25:10] What are we missing in the conversation?

[00:25:12] Right.

[00:25:12] Literally put a copilot in there, uh, in the management team.

[00:25:15] Uh, others are using it for first quick drafts, right?

[00:25:18] Going into thinking of their annual reports or engaging with the board, et cetera.

[00:25:23] Or if they need to do a quick analysis, uh, they can ask it for the first view.

[00:25:28] As we talked earlier, give me the first list of things I should think about.

[00:25:30] So it becomes a Kickstarter and ideation and thinking.

[00:25:34] And so first executives should be doing it, using it, uh, and getting building intuition around it themselves.

[00:25:39] Two, um, executives really need to make sure that the underlying infrastructure and access to the right data is in place.

[00:25:48] So that, um, they make sure that there are the right guardrails, but also access to all the company's data is in place.

[00:25:56] Cause in the end, if the tools get widely accessible to everyone and both you and I have this access to the same tool, it will be difference will come from.

[00:26:05] What is the data I can uniquely use with this tool, right?

[00:26:08] That others don't.

[00:26:09] And two, do I have the right intuition to do something with it and act on it?

[00:26:13] Right.

[00:26:13] So they got to build those two muscles.

[00:26:15] And then the third thing they need to do is build the organizational muscle to make those changes on the ground.

[00:26:21] So if we cannot do stuff faster in a different way, we'll change the incentives, change how people get rewarded to do things.

[00:26:28] But that is not an easy muscle to change.

[00:26:30] And it takes a while.

[00:26:31] Right.

[00:26:31] And, um, those are the three things executives need to do.

[00:26:34] Um, that again, as I mentioned earlier, the muscle to do that has atrophied through time to do it at this level of destruction.

[00:26:40] First class advice from you there and probably a perfect moment to end on.

[00:26:45] But before I do let you go, all this talk of Gen AI, I'm going to go old school for a moment and ask you to leave a book that you would recommend for everybody listening to check out.

[00:26:55] We have an Amazon wishlist.

[00:26:56] Ask my guest to add something to that list.

[00:26:59] What would you like to add to a book wishlist?

[00:27:02] We'll go old school on you there.

[00:27:04] Um, the, the, the book that changed, um, a little bit of how I think about the world and myself and it's, uh, Irving Stone's Loose for Life, which is a book about Vincent van Gogh.

[00:27:17] Uh, and, um, I got mesmerized by reading it.

[00:27:21] Not, I started reading it because of him, but then I fell in love in the book with the character of his brother, Theo, um, who was the art, um, dealer.

[00:27:30] Who was the one that owned most of his artwork and encouraged him to do doing the work.

[00:27:36] And I found myself being more like Theo than Vincent, because I just love to find people that have passion about something and then make sure I get the, create the right conditions for them to just go and be the best at it.

[00:27:48] And then you bring it to the rest of the world.

[00:27:50] Right.

[00:27:51] Um, and so it's similar a little bit like Gen AI.

[00:27:53] A lot of the folks three years ago were the craziest on an island talking about this stuff.

[00:27:58] Right.

[00:27:59] And so I had to have their back in a pocket to make sure, don't worry, just keep doing what you're doing and the rest of it.

[00:28:04] They're going to get it.

[00:28:05] Right.

[00:28:05] And now they got it.

[00:28:06] Now these are the rock stars.

[00:28:07] Right.

[00:28:07] So it was, that was, uh, the book that impacted me a little bit through time.

[00:28:11] Oh, I love it.

[00:28:12] I'll get that added straight to our Amazon wishlist.

[00:28:15] And for anyone listening wanting to dig a little bit deeper on the report that we've talked about today, or just find out more information about you, BCG, the kind of work that you're doing, carry on this conversation.

[00:28:26] Anywhere you'd like to point everyone listening?

[00:28:56] Yeah.

[00:28:58] And hearing more about some of the, the big stats from that report, like consultants using Gen AI completed coding tasks at 86% of the benchmark set by data scientists and 10% were faster on average is incredible.

[00:29:13] But I also love going today with you a little bit beyond that, um, report itself and how it can drive productivity gains and skill expansions for workers and even transform your child's thinking.

[00:29:27] How they can think about launching their own business one day and even plan a family vacation.

[00:29:33] Literally opportunities are endless, but just thank you for sharing that with me today.

[00:29:36] Awesome.

[00:29:37] It was a pleasure.

[00:29:37] I think as we've learned from Vlad today, Gen AI is much more than just a productivity tool.

[00:29:44] It's a game changer for knowledge work, offering unprecedented opportunities to rethink workflows, develop new skills, accelerate innovation, even generate business ideas or plan the perfect family vacation.

[00:29:58] But yes, as with any transformative technology, the challenge lies in striking that right balance between immediate gains and sustainable skill building.

[00:30:09] And understanding the limitations of this technology is equally as critical too.

[00:30:14] But what are your thoughts on integrating AI into your workflows?

[00:30:18] How are you leveraging it to shape your industry?

[00:30:21] I get a lot of emails from business leaders that are unsure on what they should be doing or what problems they should be fixing, how they can be leveraging this technology.

[00:30:30] And the plan for this podcast has always been share ideas together.

[00:30:34] Let's just use this as a big brainstorming session as well.

[00:30:38] We can share them out there.

[00:30:39] We can all learn from each other.

[00:30:41] I know myself and Vlad are going to be staying in touch, as he said at the end of our conversation.

[00:30:45] Let's keep each other sharp by learning from each other.

[00:30:48] And I encourage you to join that too.

[00:30:50] So please email me, techblogwriteroutlook.com, LinkedIn, Instagram, X, at Neil C. Hughes.

[00:30:59] Let me know.

[00:31:01] But thank you for tuning in to Tech Talks Daily today.

[00:31:05] Until next time, stay curious, stay inspired.

[00:31:08] Maybe even think about coming back and joining me again tomorrow.

[00:31:12] Hopefully speak to you then.

[00:31:13] Bye for now.