Own Your Own Intelligence Before Your Vendor Learns It

Governance
Written By Christina Ellwood

Episode Summary

Speaking in September 2024, Paul Baier, chief executive of GAI Insights, makes the case that a company using someone else’s model is teaching that supplier how its business works. His example is Toys R Us, which handed its online toy business to Amazon for a few Christmases and got it back after Amazon had learned how to position and market toys; no contract was broken and decades of process knowledge changed hands anyway. He asks what the same exchange now means inside drug discovery, oil and gas refining and stock trading, and says the first job is working out which parts of a business are genuinely its own intelligence. What he recommends is not caution: he tells companies to accelerate their learning rather than their spending, to treat an AI center of excellence as a learning hub rather than a place to send the problem, and to get their boards asking whether anyone has looked at what could leak.

Key takeaways

  • His starting point is not the technology but who owns it, and he treats that as the underexamined half. Running parallel to the societal impact of AI is what he calls the unprecedented market power of the big technology firms, where he said in September 2024 that the top five or six were larger than the third and fourth economies combined

  • He puts numbers to that power and then says nothing is coming to check it. He said in September 2024 that those firms were throwing off about 480 billion dollars of free cash flow a year, with Microsoft at 2 billion a week, that there was no trust busting coming in sight, and that the technologies involved scale exponentially

  • His warning is that the thing being collected has changed, and he reaches for a pre-AI example to show it. What moves now, he says, is an understanding of processes, and his illustration is Toys R Us executives deciding nobody would buy toys on the internet

  • The point of the example is that nothing improper happened, and he says so explicitly. In three years Amazon learned how to position toys, how to sort them by age, the graphics, the communications and the coupons; no contract was broken, and decades of process knowledge had changed hands

  • He says the same question is now being asked by big companies and not just by analysts like him, about the parts of a business that are cognitively advanced. He puts it as open questions rather than as findings: what does this mean for people doing drug discovery, oil and gas refining and stock trading

  • His own framing for the problem is deliberately larger than a procurement question. Using public models makes it hard to avoid conversations with a supplier about what the roadmap should be for your own field, and in September 2024 he wanted your own intelligence, as he puts it, discussed across 194 sovereign countries, 235 million companies and 8 billion-plus citizens rather than left to regulators

  • Before he numbers anything he says the work starts with an inventory, and that some industries have already done it. Part of it, he says, is assessing what is genuinely your intelligence and your core digitally based competitive advantage; insurance companies already have advanced AI groups and have found generative AI a natural extension

  • He names the companies he thinks are most exposed and the phrase he uses for them is his own. Companies in the crucible, he said in September 2024, were those particularly at risk of revenue decline from generative AI over the following three years, and his example is law firms, most of which have never thought of themselves as a technology partner

  • The steps he numbers start here, and the first two are both in one breath. One is to have the discussion and get alignment on the delineation; two is to work out which data and which models sit behind your firewall or inside a private trust environment, and which can sit in a public cloud for cost reasons

  • His third numbered step is to ask a lot more tougher questions, and the one he supplies is about a word. Open source in a large language model is totally different from open source in Apache or Linux, where you got the full source code and every element of it

  • He expects the management problem to arrive before the technology problem is solved. He said in September 2024 that every employee would come to have a digital twin, with thirty, forty, fifty or eighty different models running, and that managing intelligence across humans and machines is a topic companies will need time even to understand

  • Asked whether companies should slow down and answer these questions first, he says the opposite, with a caveat. This is like swimming: get in the water, start at the baby pool, and accelerate your learning processes, which he is careful to distinguish from accelerating your spend

  • His cheapest suggested starting point is a real one rather than a gesture. He said in September 2024 that there was a whole slew of great products that could put a chatbot with public data on your website for under five hundred dollars and about two hours, on a knowledge base you already have

  • His analogy for organizing all this is the quality movement, and he uses it to argue against concentration. Quality needed levels of expertise, white belts and black belts, and it needed to live at the process level rather than in one team

  • His own line for the failure mode is that the work becomes his job and not my job, said of the VP of quality. He carries it straight over: an AI center of excellence read as a learning hub supporting learning across the organization is a good model, and one where all the decisions and all the learning sit in that group is not, in his view

  • Asked whether org structure changes the answer, he thinks it does not. Loosely coupled or top-down, he says it comes down to leadership: whether the leadership encourages experimentation, learning and rapid adoption. Org structure, in his view, does not solve the imperative of leadership

  • His challenge to a board is an argument by analogy, and he says he means it to be dramatic rather than controversial. He asserts it would be irresponsible for a board to hire a VP who has never used a browser, and then asks the board to accept the same standard for anyone opining on generative AI strategy who has not personally used the technology for three hours

  • His framework for boards is an acronym and he expands it on air. WINS is words, images, numbers and sounds, a subset of knowledge work, and it is meant to help an industry, a company or a job work out whether this is urgent for them

  • What he actually wants from a board is a question rather than a decision. How are we thinking about our cognitive assets, do we even know what they are, how much can be automated, how much can leak out, and he says in most cases that conversation has not started

  • His test for whether the technology has landed with someone is a phrase they start using. Hey, have you asked your robot, which he calls a good question and treats, in his view, as the critical path in moving this from parlor trick and novelty to indispensable tool

  • His closing recommendation replaces one-off reengineering with something continuous. Take every process down into subtasks and ask of each one whether the human or the robot leads, then accept that the answer may reverse six months later

About Paul Baier

Paul Baier is chief executive and a co-founder of GAI Insights, which the host introduces on this episode as the leading GenAI analyst firm. He co-wrote the Own Your Own Intelligence framework this episode is built around with John Sviokla, who is the guest on episode 34. His argument here is that the market power of a handful of large technology firms has changed what it means to buy software: a supplier processing your data learns how your business works, and that knowledge does not come back. He ran learning groups for member companies and advised boards at the time of this conversation in September 2024, and he is more interested in getting executives to use the technology personally than in getting them to read about it.

 

In this episode

00:35 Welcome, and who Paul Baier is
00:58 The question: what own your own intelligence means
01:23 Market power as the parallel story
01:46 Free cash flow, and no trust busting in sight
02:16 Toys R Us decides nobody buys toys online
02:48 What Amazon learned, and what no contract covered
03:12 Drug discovery, refining and stock trading
03:32 A conversation for countries, companies and citizens
04:11 The question: practical steps
04:34 Before the list: assess what your intelligence is
04:56 Companies in the crucible, and law firms
05:14 His steps one and two: alignment, then what sits behind the firewall
05:35 His step three: ask harder questions about open source
06:31 A digital twin for every employee
07:14 The question: slow down, or start now
07:29 Not slowing down: this is like swimming
07:34 Get in the water, and accelerate learning not spend
07:54 A chatbot for under five hundred dollars
08:31 The question: is that a fair summary
08:58 The question: an AI center of excellence
09:19 The quality movement, white belts and black belts
10:02 Learning hub or outsourced destination
10:22 The question: does org structure matter
10:38 It comes down to leadership
11:03 The question: advising a board
11:11 The VP who has never used a browser
12:20 The WINS framework
12:58 The question: should the board lead this
13:33 The board is there to ask good questions
13:42 Cognitive assets, and what can leak out
14:22 The AGI question he says boards should at least ask
15:05 The New York Times story he calls breathtaking
15:39 The question: where to learn
16:00 Living in a hurricane of change
16:47 Have you asked your robot
17:39 The question: the closing takeaway
18:19 Every process into subtasks, human or robot

In Paul’s words

“There is no trust busting coming in sight, and they are on a roll”

Paul Baier   (01:46)

“every one of our employees is gonna have a digital twin”

Paul Baier   (06:31)

“You have to get in the water and learn, and you’ve got to start at the baby pool”

Paul Baier   (07:34)

“It’s his job and not my job.”

Paul Baier   (10:02)

“the key question for the board is ask good questions”

Paul Baier   (13:33)

“how are we thinking about our cognitive assets? Do we even know what they are?”

Paul Baier   (13:42)

“we kinda gonna live in a hurricane of change”

Paul Baier   (16:00)

“parlor trick and novelty to indispensable tool”

Paul Baier   (16:47)

 

Resources

Named on air

  • Toys R Us and Amazon: His worked example from 02:16 to 03:12, of process knowledge transferring to a partner without any contract being broken

  • Llama 3: Named at 05:55 as his example of a model where the weights are published and the training data is not

  • Apache and Linux: Named at 05:35 as what open source used to mean, where you got the full source and every element of it

  • The New York Times: His attribution at 15:05 for the story he calls breathtaking, about a man with ALS able to speak with his family again

  • The GAI Insights weekly learning lab: Named at 16:28. He said in September 2024 that three thousand members took part

Ideas and terms discussed

  • Own your own intelligence: His framework: a call for countries, companies and individuals to keep control of the knowledge that makes them distinctive

  • Companies in the crucible: His term at 04:56 for firms he said in September 2024 were particularly at risk of revenue decline over the following three years, where law firms are his example

  • Free source models: His phrase at 05:55 for models published as open source where the weights are available and the training data is not

  • The WINS framework: Words, images, numbers and sounds. His tool at 12:20 for judging how urgent this is for a given industry, company or job

  • Cognitive assets: His term at 13:42 for what a board should be asking about: what they are, how much can be automated, and how much can leak out

  • Digital twin: His September 2024 forecast, at 06:31, that every employee would come to have one, alongside thirty to eighty models running in the organization

Related AI Realized episodes and events

 

Frequently Asked Questions

  • Owning your own intelligence means keeping control of the knowledge that makes your business distinctive, rather than handing it to a supplier that processes your data. Paul Baier of GAI Insights, speaking in September 2024, frames it as a question to answer before a purchase: which parts of this company are genuinely its own intelligence and its core digitally based competitive advantage. He wants that treated as a discussion across countries, companies and individuals rather than as a procurement detail, and he is explicit that he does not expect regulation or open source models to settle it.

    Transcript 00:58 to 04:34


  • A supplier learns your business by doing the work, which no confidentiality clause covers. The example Paul Baier of GAI Insights gives is Toys R Us, which decided nobody would buy toys online and let Amazon run the category for a few Christmases; when it took the business back, Amazon had learned how to position toys, sort them by age, and handle the graphics, the communications and the coupons, and went into the toy business itself. Nothing in that broke a contract, and decades of process knowledge had changed hands.

    Transcript 02:16 to


  • The ones whose product is cognitive work and who have never thought of themselves as technology companies. Paul Baier of GAI Insights calls them companies in the crucible, particularly at risk of revenue decline over the following three years, and names law firms as his example. He contrasts them with insurance companies, which already had advanced AI groups doing traditional AI and found generative AI a natural extension, and he points at drug discovery, oil and gas refining and stock trading as the other places where the knowledge at stake is the business.

    Transcript 03:12 to 05:14


  • No, and the advice runs the other way: accelerate the learning. Paul Baier of GAI Insights is careful to add that this does not necessarily mean accelerating the spend, and compares the whole thing to swimming: you have to get in the water, and you start at the baby pool. He said in September 2024 that products then available could put a chatbot with public data on a website for under five hundred dollars and about two hours, which he treats as a way to learn rather than as a deployment, and he adds that the people in an organization who take to it first will show leadership where the real opportunities are.

    Transcript 07:14 to 08:31


  • It depends on whether the rest of the organization reads it as a resource or as a place to send the problem. Paul Baier of GAI Insights says he has no strong recommendation on the right organizational mechanism, and draws a parallel with the quality movement instead, which needed levels of expertise and needed to operate at the process level, and with the companies that appointed a VP of quality and thereby told everyone else it was not their job. A center understood as a learning hub supporting learning across the organization is a good model; one where all the decisions and all the learning sit in a single group is not, in his view.

    Transcript 08:58 to 10:22


  • A board should be asking whether anyone has identified the company’s cognitive assets and what could leak out of them. Paul Baier of GAI Insights puts the board’s job as advising, hiring and firing the chief executive and asking good questions rather than making the calls itself, and the questions he suggests are open ones: how are we thinking about our cognitive assets, do we even know what they are, how much of them can be automated, and how much can leak. He said in September 2024 that in most cases that conversation had not started.

    Transcript 12:58 to 14:22


  • Yes, and the case for it is made by analogy rather than by assertion. Paul Baier of GAI Insights says it would be irresponsible for a board to hire a VP who has never used a browser, and asks boards to hold anyone opining on generative AI strategy to the same standard: at least three hours of personal use. His reason is that the experience does not transfer through reading, and his comparison is travel, where you cannot research your way to what a country is like.

    Transcript 11:03 to 12:20


  • WINS stands for words, images, numbers and sounds, and it is a way of asking how exposed a particular job or industry is. Paul Baier of GAI Insights describes it as a subset of knowledge work and as a handy framework for helping industries, companies and jobs understand urgency, with the explicit caveat that not everything is urgent for all companies. What he cares about is that a board reaches a shared answer: he says the alignment matters more than any individual member’s view, and that without an explicit conversation you get neither alignment nor sensible investment.

    Transcript 12:20 to 12:58


 

  • [00:35] Christina Ellwood: Welcome to AI Realized podcast for enterprise executives adopting AI. I’m Christina Ellwood, your host for today’s episode. We’re talking today with Paul Baier, CEO of GAI Insights, the leading GenAI analyst firm. Paul, welcome to AI Realized.

    [00:54] Paul Baier: Thank you, Christina. Excited to be here

    [00:56] Christina Ellwood: Well, we’re really glad to have you, too. AI depends, of course, upon vast amounts of data, and it learns our behaviors as it goes along in learning with us. So in your recent HBR article, you and John made a call for countries, companies, and individuals to, in your words, “own your own intelligence.” Can you elaborate on what you mean by this?

    [01:23] Paul Baier: Absolutely. We are inflection as a society in many different areas, one of which is trying to understand the societal impact of AI technologies. But parallel to that is also understanding the unprecedented market power of the big tech firms. The big tech firms, the fi- top five or six are larger than the third and fourth economy combined. They’re kicking off about $480 billion of free cash flow a year. Microsoft kicks off $2 billion a week. There is no trust busting coming in sight, and they are on a roll, and they’re on a roll with, uh, ex- the technologies that scale exponentially. Why does that matter? We have seen in the last 10 to 15 years the location data and metadata of all the citizens and the companies in the world used for competitive advantage by the high-tech firms. We are now entering a world where there’s even more data that’s out there in terms of understanding processes. And the simple example I’d like to share is what happened with, uh, Toys “R” Us in the early days of the internet. In the early days of the internet, Toys “R” Us executives said, “No one’s gonna buy toys on the internet. That’s a stupid little novelty item.” So they said, “Amazon, why don’t you do the toys for us for the first couple Christmases?” Which they did, and they signed a contract. After three years, Toys “R” Us figured out, “This is something important. We should probably do this ourselves.” And Amazon said, “Fine, you do it yourselves. We’re gonna take what we learned the last three years and now be in the toy business.” In those three years, Amazon learned things like how to position the toys, how to, um, position it by age, the graphics, the communications, the, uh, coupons here. None of that was, uh, contractually breaking contracts here, but that process knowledge now had been transferred from decades from, uh, Toys “R” Us now to, uh, Amazon. And the same question is being asked by big companies, not just us. They’re being asked by companies, and we’re reflecting what we’re seeing out there. It is, what does that mean for real important parts of my businesses that are cognitively advanced and now using AI? What does it mean for people who are doing drug discovery and oil and gas refinery and stock trading? And when I’m using public models, it’s hard for me not to have, at some point, a conversation with, let’s say, the product manager talking about what the roadmap should be for drug discovery Because we want that vendor as in traditional stuff to use it. And your own intelligence is a framework, a global conversation that we wanna start where the 194 sovereign countries, the 235 million companies, and the 8 billion-plus citizens are starting to grasp with this because it’s gonna take more than just hoping for regulation out of Seoul and the EU and open source models to, um, have a fair and equity distribution of this power structure.

    [04:11] Christina Ellwood: So it sounds like as much as we have the opportunity to innovate and redesign our businesses and create a competitive edge, this technology also represents a threat, a competitive threat, an existential threat. So what are some of the practical steps that organizations need to take to implement your guidance on own your own intelligence?

    [04:34] Paul Baier: I think part of it is starting to assess what really is our intelligence and what really is our core competitive advantage here that’s digitally based. Many companies have already been doing this for years. Insurance companies, for instance, already have very advanced AI groups that do use traditional AI, and gen AI has been a natural extension of that, and that’s been more natural for them to think about this. Other companies, which we call in the crucible, which they’re particularly at risk of revenue decline because of gen AI over the next three years, aren’t, and a good example of that’s law firms. There’s unbelievable transformation coming to law firms. Most law firms have never thought of themselves as a technology partner here. Do they need to own AI? Do they need to develop into models here? So the practical steps we recommend are as follows: One, start having this discussion to get alignment on the delineation. Two, start understanding options of maybe some of the data and some of the models are behind your firewall or behind your trust environment in a private cloud. Others are maybe in the public cloud for cost-effective reasons for a chatbot on your website, for instance. You know, third, start asking a lot more tougher questions. Open source LLM is totally different than open source Apache and Linux. Those open source things, you actually got the full source code and all the elements of it. That’s not true today with the, quote, “open source model.” They’re almost like free source models. You might get the weights in Llama three’s cases, but you don’t get the training data. You don’t get access to, uh, some of the, um, ways to actually run the model here. So it really is a difficult and important conversation. And then the third area the companies to start thinking about is what really does my workforce look like five years from now, and how do we manage it? So if I have a five thousand person organization today, I have five thousand person organization, I have workforce management, talent management for humans. We are moving towards a world where every one of our employees is gonna have a digital twin, and we’re gonna have thirty, forty, fifty, eighty different LLMs running out, and I’m gonna actually be managing my intelligence for my most important value creation processes across a collection of humans and AI-based or robots and humans. And how do I think about that, and what are the boundaries and what are the decision rights is a really important topic that’s gonna take a while for companies to even understand the issues, much less before they start implementing on practical solutions.

    [07:02] Christina Ellwood: Sounds like it could become a barrier to adoption if we don’t have answers to those questions, how do we manage the data and the security and these intelligence assets? Do you feel like companies need to slow down and answer the questions before they begin, or do you feel like there is a way that they can get started and figure out their intelligence assets and how to protect them longer term over a longer timeline?

    [07:29] Paul Baier: It’s a great question. I think the thing that we see most is that this is like swimming. You have to get in the water and learn, and you’ve got to start at the baby pool and go up here. So we encourage companies to do whatever you can to accelerate your learning processes. That does not necessarily mean accelerate your spend. There’s lots of things that companies can do for no code, uh, large language model products, for instance. There’s a whole slew of great products out there from, uh, ChatHub or, um, custom GPT, AI, and others, where you can have your own chatbot with... on your website with public data for less than five hundred dollars and two hours, and start taking and putting in LLM-based, a chatbot in every single one of your knowledge bases. On your public site, for your product descriptions, for your onboarding employee here. That’s a great way to start learning. And when you start learning with your AI rebels and your AI enthusiasts, they will then take you to the new opportunities in the organization here, and then that will help executives understand it here.

    [08:31] Christina Ellwood: So that sounds like start with your public data and get your feet wet and begin to learn and gain the skills and the-- understand the opportunities, and move into using your proprietary data and find your intelligence assets as you’re doing both of those steps. Is that a fair summary?

    [08:51] Paul Baier: That’s correct. And, and the whole time learning and learning as a cross-functional team and not just within an IT silo.

    [08:58] Christina Ellwood: Okay. And in fact, I was going to ask you if you were recommending to people that they form an AI center of excellence or some equivalent kind of a designated group within the enterprise to take responsibility to concentrate the initiatives and the learnings for the organization.

    [09:19] Paul Baier: We think that a lot of, uh, a good analogy to learn through this is look at the quality movement here. And the quality movement was a set of processes and thinking that needed to be able to drive value, you needed to have levels of expertise, white belts and black belts, and it needed to be at the process level. It didn’t mean you weren’t rigorous, but it means you needed to get that organizational thinking down here. So that’s the goal. I think we don’t have a strong recommendation exactly what the right organizational mechanism is, other than the fact that if you create a center, you have to make sure you don’t send a message to the organization, you don’t need to worry about it here. If you remember back in the early days of the quality movement, many, many companies had a VP of quality, and therefore the rest of the organization said, “I don’t have to worry about quality because we have a VP of quality. It’s his job and not my job.” And it’s the same thing on AI center of excellence. If the AI center of excellence is seen as a learning hub to support lots of organizational learning, that’s a great model. If it seem like, oh, it’s an outsourced thing, all decisions and all learning is in that one group, then that’s not an effective model in our view.

    [10:22] Christina Ellwood: Do you think it matters whether the organization is a complex, loosely coupled set of multiple players that like you have in a healthcare system or a country versus a company where everything is under the command and control of a single management team and board?

    [10:38] Paul Baier: I think like in a lot of things, it comes down to leadership. You know, do you have a leadership in either loosely controlled or, you know, top-down that’s encouraging experimentation, learning, rapid adoption of new technologies or not? And I think those cultural issues or policy issues are above and will perva- be pervasive in whatever the org structure is. So org structure doesn’t solve the imperative of leadership in our view

    [11:03] Christina Ellwood: If you were talking with a board of directors, what would you be advising them on their role in leading in this area?

    [11:11] Paul Baier: Well, first of all, I would congratulate them at least having a meeting at the board level on gen AI, ’cause I would be there, I’m assuming. We’ve done a lot of board meetings here. The second thing I go down, and it’s not meant to be controversial, it’s meant to be, um, dramatic, is that I assert that it would be irresponsible for this board, and they-- see if they would agree, to hire a VP in their organization who’s never used a browser, ever. Everyone says, “Let’s agree here.” And I assert in this age of AI, it’s equally responsible to have a board member or senior management member who opines on gen AI strategy or capital X or capital allocations who hasn’t personally used this technology for three hours. It is easy to use. If you’re one of the eight point one billion people that use Google and you know how to type something in a box, then you can type something in a box. You’re a prompt engineer, you can put in the context window and try it on a ChatGPT and start experiencing the transformational element, ’cause this is different. You can’t research your way to what China or Japan’s like. You need to go visit it, and then you understand better, and it’s the same thing for that. So that’s the first set of things. And the third thing I-- we recommend to boards is to use our WINS framework. WINS stands for words, images, numbers, and sounds. It’s a subset of knowledge work, and it’s a handy little framework to help industries, companies, and jobs understand the urgency. Not everything’s urgent for all companies, but it is for some. And it’s less important what any individual board member’s view on it. It’s much more important what the alignment is at that board level. Is it urgent for us at this particular company at this particular moment? And without explicit conversations on that, in our experience, you aren’t getting any kind of alignment, much less optimized investment.

    [12:58] Christina Ellwood: Do you think the board needs to lead the conversation about for the own your own intelligence risk factors? Like, you know, do we understand our intelligence? Have we, you know, addressed and mitigated the risk? Do we need to have a designated committee related to it? Is there anything like that, that you think the board needs to consider, especially for companies that have so much process intelligence that would potentially be leaked? Yeah. Much like you talked about the Toys “R” Us case, but where you would have behavioral and process leaks through the use of these tools.

    [13:33] Paul Baier: I think a lot of things, that the board’s there to advise and hire and fire the CEO and a few other administrative things here. So I think the key question for the board is ask good questions. And when board members ask us what questions to ask around this, we really encourage them to ask obviously open-ended questions, but questions around how are we thinking about our cognitive assets? Do we even know what they are? You know, how much of those can be automated? How much of those can leak out? And in most cases, we have found that conversation hasn’t even started. So part of it is educating the board on the realities of the market power out there. The second is to help the board understand how fast things are changing. And third is to start figuring out what are questions to ask and get answers from, or at least start the dialogue about, is there really ways to have process leakage here? Is the .002% or maybe 4% chance in the next 10 years that Microsoft gets true AGI, does that represent a risk? They have all our data. Does that matter? And at least start asking questions to, uh, start having meaningful conversation on that, because right now that conversation’s not happening in most boardrooms.

    [14:42] Christina Ellwood: Yeah, I think those are some of the same questions that need to be asked at the management team level.

    [14:42] Paul Baier: Yes.

    [14:42] Christina Ellwood: So yeah, it makes good sense. Just shifting gears a little bit into the analyst view that you hold. You know, clearly AI can be used to optimize existing processes, but it can also be used to do things that were previously impossible. Which do you predict will make the biggest impact on the world?

    [15:05] Paul Baier: Oh, I, I think the previously impossible is, we’re just scratching the surface. I mean, the recent New York Times article about the, uh, gentleman who had, um, ALS or Lou Gehrig’s disease, and now is able to talk to his family again. I mean, that’s breathtaking, and that was beyond the boundary condition of what people thought possible 12 months ago. And we’re gonna continue to have, in our view, dozens and dozens, and dozens, and dozens of examples like that. So on the AI optimistic side, we could be looking at a transformational world just 10 years from now

    [15:39] Christina Ellwood: That’s a very exciting prospect, but a little unsettling when we have so many questions that are unanswered, aren’t they? What do you advise for people who want to learn more and tap into resources that can help them with their own framing and decision-making related to adopting this technology?

    [16:00] Paul Baier: It’s, um, really important, I think the first and foremost, is to understand the, embrace the pace, understand the change is not gonna slow down, so we’ve gotta get used to this wind, so we kinda gonna live in a hurricane of change. And then secondly is figure out, for your personal learning style, what works. How do you create your own learning group? Do you learn by reading? Do you learn by doing? Do you learn by teaching? Learn by, you know, reading and then doing? And then how do you find a learning group, a learning partner that works? We have a weekly learning lab that’s 3,000 of our members are part of, that’s been very effective for them. We do a lot of learning groups within companies. We encourage them do it. We know people that are partnering. But finding someone else to, quote, “learn to swim,” finding someone else to learn to be constantly using the robots throughout the day. “Hey, have you asked your robot?” That’s a good question. Have we asked the, um, our co-intelligence? Is the critical path, in our view, of moving this technology through the human change management process from parlor trick and novelty to indispensable tool, which is quickly moving towards, and people who get there fast enough will see the biggest advantage, both personally for their career and for their companies

    [17:10] Christina Ellwood: I’m personally very impressed with the commitment and the follow-through that your organization has brought to the marketplace for people who wanna learn about AI. Not only do you do the learning labs and your emails, you are incredibly prolific in your writing. There’s four different HBR articles that were published recently, and you’re very generous in sharing your thoughts and your knowledge and expertise, and I just wanna, you know, give you a shout-out that I’m, I really respect what you guys are doing. But as we wrap up, what would you like our listeners to take away from our conversation today?

    [17:46] Paul Baier: Christina, thank you for the kind words of the whole team here, so I can’t take a lot of credit for that, but, uh, it really is appreciated. And then we are very open and hope to, uh, continue to inspire others to do the same thing ’cause we need a lot of learning here. I want people to take away that this is not changing, so we need to completely reorient our mindset, figure out what’s in our control, and what’s in our control is supporting each other, learning, and understanding quickly here. We’re gonna have to keep learning and relearning tools. It’s gonna be change, and that’s why one of the articles we wrote was designed with dialogue, where how you... instead of the traditional process reengineering, you took one process, you reengineer it, and it was fine for the next five years. Here, we need to take down every process and can break it down into subtask, and for each subtask, ask, “Who does the lead? Is it the human or the robot? Human or the robot?” And then know over time that sometimes a human’s gonna be leading, but maybe six months from now, the robots will be leading here. So that kind of different framework on how we think about change, different framework on how we think about who does work, and different framework on how we think about managing risk is the most powerful mindset and capability developing strategy there is for sanity as well as, uh, career advancements over the next five years.

    [19:03] Christina Ellwood: Well, Paul, thank you so much for talking with us today on AI Realized. We really appreciate you taking the time and sharing your learnings with us.

    [19:12] Paul Baier: Thank you, Christina.


 
Listen to other episodes
 
 
 
Join AI Realized
 
 
 
Subscribe to the newsletter
 
 
 
Christina Ellwood
Previous

Bring the AI to Your Data, Not Your Data to the Cloud