How do you build skills in the age of autonomy?

FRED VOGELSTEIN:

00:00:00:01 Let's go. This morning I watched an old Twilight Zone episode.

APARNA CHENNAPRAGADA:

00:00:08:09 As one does.

FRED VOGELSTEIN:

00:00:09:12 As one does. it was actually the story about how, CEO WUV Whipple decided to introduce computers into his factories and how excited he is about how many people and costs his machines can eliminate. Until those computers end up eliminating his job too, so it's actually a little chilling. But also interesting for our purposes today because it aired in May of 1964.

00:00:44:11 61 years ago. And it actually turned out to be completely wrong. The digital revolution of the past 75 years, has changed and eliminated a bunch of jobs, but it's also created, a lot more new ones. But AI is going to change how we work. It will invent jobs that don't yet exist. And the transition has the potential to be bumpy.

00:01:15:16 I think many of you may have seen the New York Times piece yesterday about how Amazon thinks that it can use AI and robots to eliminate half a million jobs by 2030. That's a really big number. So, over the next 50 minutes, we're going to talk through how AI is changing work and what skills we're going to all need to live in that world and which ones will be obsolete.

00:01:47:14  So my goal is for us to have a conversation for the next 35 minutes. And then leave about 15 minutes for questions, because this is a wonderfully big and interactive audience. I can just tell. So, to start, I just want to let all of our panelists introduce themselves. And then after that, I'm going to ask each of them to tell us a little bit about what they're seeing.  Go ahead, Anthony.

ANTHONY ABBATIELLO:

00:02:18:09 Hi, good afternoon. It's like a warm room that everyone's standing. And so feel free to move around. Anthony Abbatiello, Partner at PwC, and I'm our Future of Work Leader. So everything we do around the work, the workforce, the work if you think about everything that's happening in the age of AI, really,

00:02:38:05 and what's happening with humans and what the future of the workforce will look like. I've spent 30 years of my career all in consulting around talent, leadership and HR function. So, it's a personal passion of mine. I've lived through so many of these industrial revolutions. I am both a student of a researcher and a practitioner in the space.

DIYA JOLLY:

00:02:59:09 Diya Jolly, I am the Chief Product and Technology Officer at Xero for those of you who don't know, Xero provides small businesses with finance and accounting and cash-flow software, so that they can actually do what it takes to run their business. We have been heavily investing in AI and I'll talk a little bit about that to actually help our customers really be able to make their work more enjoyable and focus on where they need to do to grow their businesses.

00:03:42:07 Prior to Xero, I was the Chief Product Officer at Okta and before that I was at YouTube, running YouTube monetization.

ELENA SUNSHINE:

00:03:53:18 Hi, everybody. My name is Elena Sunshine. I'm thrilled to be joining this panel and being among such esteemed colleagues here. I am a Senior Director of Product Management at Oracle and I focus specifically on our OCI, that’s our cloud service for building generative AI platform that includes our agents offerings and also our AI safety offerings.

00:04:16:16 I work a lot with large enterprises who are approaching their businesses and looking at how to use AI to solve their problems and our customers really care about cost, security, privacy, compliance, reliability. So that's kind of the key focuses that I have in my day-to-day life.

00:04:40:10 And I also work a lot with the entire stack of Oracle applications across HR applications, finance, supply chain management, ERP, the gamut to integrate AI into those technology stacks as well.  I'm thrilled to be here. Thank you.

APARNA CHENNAPRAGADA:

00:04:58:00 Hey, everyone. Glad to be here. Yes, it is warm. My name is Aparna. My last name is Chennapragada, which I think in solidarity I'm going to say in my language, it means happiness because you're Sunshine and Jolly — I’m making it up.

ANTHONY ABBATIELLO:

00:05:12:13 I was like, oh, my God, I have to make something up now. I'm very happy Italian.

APARNA CHENNAPRAGADA:

00:05:20:17 I am the Chief Product Officer for AI at work at Microsoft. So what I look at in our team's work on is how does AI change and amplify how we work, how we create, how we collaborate, and how we live. So, this is very topical. I have some thoughts that I'd love to kind of share and get some questions and so on.

00:05:42:11 I've been working in AI space for the last two decades. I'd say I worked on Google search, built a couple of new products, Google Assistant, as well as Google Lens, which is more of AI product to search what you see. And all through, I think one passion and one steel thread for me is how do we actually build on new, deep technological insights, but build them in a very intuitive way in a positive some way.

FRED VOGELSTEIN:

00:06:09:20 Aparna, two decades in AI — that makes you one of a very small group. Can you talk a little bit about what you're seeing? I mean, there seems to be a conversation going on, right as we're speaking about. Is AI actually helping productivity or is it hurting productivity? And, what are sort of some of the best practices that are starting to develop?

APARNA CHENNAPRAGADA:

00:06:39:11 I'll try to condense it into a few thoughts. Obviously, this is a topic that lives rent-free in my brain. So, if folks have questions about it, we should talk later. I'd say, look, I've been in the forefront of internet shift. I was at Akamai, building the first content delivery system, and then the mobile and like, web shift, obviously.

00:06:59:07 And now and folks often say, well, is this different? This time is different. That phrase keeps coming up. And at first, I was like not really like every technology shift has the first phase where there's an exuberance, there's an overbuilding, optical networking. Those of you who remember 2000s — significant amount of money and like build out in infrastructure.

00:07:21:03 The use cases came much later and so on, right. But I will tell you, number one, this time is different in a very particular way. The technology diffusion usually and Anthony, you've seen it too which is it used to take decades from the first time, I don't know, this giant brick of a mobile phone was in mid 90s to like, I don't know my uncle in like small town in India getting a phone that's like a good 20 years.

00:07:47:14 Now, it's months. In fact, as we speak, I'm sure there's a model being dropped somewhere, right? So, it is one of those motion sicknesses inducing real compression of technology shift that's happening. The corollary and what it's relevant to what we are talking about here, is that the adoption, the user adoption of this is significantly higher in a small setting.

00:08:12:08 Now that doesn't mean it's evenly distributed. That's why you see the paradoxical, oh, these pilots don't work. And yet everybody has shadow usage in the enterprise.

FRED VOGELSTEIN:

00:08:23:16 Why don't you go ahead Anthony?

ANTHONY ABBATIELLO:

0:08:25.16 I agree. I think what we're experiencing right now is this movement, to the point of talent now on one hand is divide. The younger talent that are AI natives that have grown up around technology. They see this as it's just table stakes.

00:08:46:18 We have to have AI, we have to have technology as part of this. Those who are in older generations. I'll put myself in that as a Gen-Xer, our generations and the baby boomers have seen this as oh, it's robots coming to kill us and take our jobs. It's the Twilight Zone example.

00:09:00:06 And so, what's happening, though, is because of the pace that we're seeing, there is no ability to sit back and watch it and see it destroy itself because it's actually not. And it's proving itself every day. When we produced a piece of research back in January, if anyone was in the earlier session with Dan Priest, we talked about that.

00:09:23:04 What was happening in January. We're looking at the future of work. Here we are in November. I almost have to rip up the entire piece of research, because we will learn so much more. And the thing that has remained constant is that the human skill is still paramount to the technology and what is changing in the environment.

00:09:42:14 My job is to only think about humans and what skills will be in the future. And nothing has proven that jobs are being taken away. I loved yesterday somebody said AI is not going to take your job, but talent with AI skills will take your job. That's the key piece that continues to resonate with me and I try to impart that on so many organizations, big and small, because the pace, it will continue to outpace anything we've ever seen before.

00:10:11:23 The models will continue to learn faster and will create new things. And we have to remember that in all parts of time, particularly like The New York Times article, I can debate ad nauseam. But I think at all moments in time when we’ve seen another revolution, it is only created new opportunity that has pushed society forward.

00:10:35:10 And this one is one that actually will perpetuate all levels of society, not just the working adult or the worker in general. It's affecting now down to children, through education, through to the experienced worker.

FRED VOGELSTEIN:

0:10:51.31  Go ahead. What I was going to ask is that, like, you spend a lot of time thinking about the small business world and I sometimes think that, the way they think about AI has got to be a little bit different than the way giant enterprises do. Maybe more efficient.

DIVA JOLLY:

0:11:10.11 Absolutely. So, I was going to bring in a different perspective and disagree a little bit. For what we see, we sit in the Bay area or in tech corporations or in large corporations, but most of the world doesn't sit there. A lot of the world is like small businesses. So, one of the constant challenges my marketing department has is how do you explain what agents are and what it can do for you?

00:11:35:17 Because they're like, what are you talking about? So, going back to your original question, here's what we see.  AI is new, it's like it's two years old or Gen AI is, LLMs are two years like old in mass, right. And no, they cannot do everything that people are saying that they can do. That's very true. However, to get adoption, you really, truly, for the user has to give value.

00:12:00:16 And there are two types of value, value that fits into your workflow. I'll give you an example. We remove 22 hours of a small business’s month of work, with one of our features that we call automated bank rec. It just basically does a large part of your bookkeeping for you. And it was right in their workflow and they adopted it.

00:12:20:16 There are different types of things you can do with AI and we're struggling with this internally at Xero, which requires a complete rethink of your workflow for AI to be effective. We're exploring AI right now in how much can we give an actual AI agent? How much of software can we have an actual AI agent?

00:12:39:19 And we have an old tech stack, right. That actually requires the way our software engineers — their entire software development lifecycle. And that is actually harder. So I think places you're having success with AI today is where there's concrete value.

00:12:58:10 It's in the workflow, right? And it actually makes your workflow faster, better and gives you time back and places where either the use cases are stretched or not proven or cases where you have to change the workflow. But people are not changing the workflow, but they're trying to slam it into an existing workflow is where I think you see the pilots struggle a lot.

FRED VOGELSTEIN:

00:13:18:15  Right. Go ahead, Elena.

ELENA SUNSHINE:

00:13:19:23 Well, I come from an enterprise context, but I'm seeing a lot of the same things, but with a slightly different perspective. So, I think the most common question that I got two years ago are what are the use cases for AI? And I can't tell you how many PowerPoints that we worked on that had lists of use cases for AI.

00:13:39:19 And that was kind of the deliverable that all of our customers wanted to show to their board. Like, here are the AI use cases that I'm going to implement. And we're going mostly in cost cutting kind of fashion. And I have seen a lot of transformation of that over the past years, I think

00:13:56:02 even within our own company, the place where I've seen AI impact work and actually really drive outcomes is by putting the tools into our employees’ hands directly rather than, I mean, I'm a product manager, so I want to build software and deliver it to someone.

00:14:14:22 But, I've really more shifted my focus towards building platforms where our employees and users can actually build themselves. I think collectively we're all still figuring out what these technologies can do. And I think when you put the tools, powerful tools into employees’ hands, they build creative and surprising solutions to their real problems, their actual business problems, rather than what you might think would be an AI problem on a PowerPoint slide.

FRED VOGELSTEIN:

00:14:45:21 Go ahead, Aparna.

APARNA CHENNAPRAGADA:

00:14:46:14 I think it's interesting, actually, this phase is very consistent across different enterprises, even on consumer. And I would argue it's not that different from like, say, the previous shifts. It used to be any new technology would have to just add X phase and then you have the X native like just add AI is the phase that we are seeing Chat app here and chat app here and meeting users where they are in their workflow.

00:15:10:20 You get incremental productivity improvements there. But I would also urge us to think about if especially the builders in this group, it's incremental productivity gains with the just add AI phase. But now the models are getting significantly better. Now they can reason, they can run for a longer time. So really rethinking entire workflows while it's harder, I actually think that that has a instead of X percent improvement.

00:15:36:03 We are seeing X times. And I wanted to give like a real example from this morning when I was in a meeting with a CIO of a company where they're adopting some of the agents that we built. One of the agents is called researcher. And the idea was to say, look, if you're a salesperson preparing for a prospective customer meeting.

00:15:56:12 Today you go through all of the world's information, like, look up, search engines, look up emails, look up meeting transcripts, and then you go and say, how do I close it.  What this agent does is, it's not trying to save you time and summarize your email or what have you. It covers the entire web.

00:16:15:03 It covers all of your meeting transcripts, all of the emails, and not only your memory, but the institutional memory about this account and says, here's the best way to close the deal and coach you on that. That's an example of, again, like it's not like an X percent. It's giving you superpowers versus saving time.

00:16:32:01 And if you're building something in AI, I would say yes, of course, do the incremental stuff because then it's you get quick ROI, but don't sleep on the kind of the rethinks.

FRED VOGELSTEIN:

00:16:42:13 Like CRM on steroids, almost. One of the questions I was, I've been sort of pondering is that, like the technology revolution of the 1980s and 1990s was largely an enterprise-led revolution that consumers then adopted subsequently, the technology revolution that started really in the 2000s with the iPod and the iPhone.

00:17:12:18 And the explosion of Google search was really a consumer- led revolution then oddly, enterprises followed, which is this?

ANTHONY ABBATIELLO:

00:17:28:02  Well, I think right now that this is the agent-led revolution that is taking this and that's not to say that it's robotic, but when I look at what is actually happening, it is coming from both angles. You're seeing it from the (Unintel phrase _____17.52) side. We're seeing it.

00:17:52:18 People are using ChatGPT.  My father is retired and he's downloaded ChatGPT and is using it, something like that. That's a win. And then, we see, kids are using it now. It's only the way in school that it's accepted now, ways of studying and learning and producing content.

00:18:10:19 So they have to use their brain in different ways, more creatively versus rote memorization. So, it could be from both angles. And then we're seeing enterprises, companies that are creating this. And it's coming at all angles. But the common thread throughout all of that is the agentic revolution.

00:18:28:14 And to me, I think that's the part that what's so interesting about this, because of the way it's learning, what we're not seeing is task and process automation, which is large part of what we saw in the 80s, 90s and early 2000s.

00:18:39:17 It was like, just take this process and make it now happen through a client server app or a cloud-based app, SaaS. And we just changed the underlying technology. Now, with the models the way they are, it's learning faster and creating new ways of thinking and actually performing the process. So now we're getting to a whole different way of thinking about, not just automation, but autonomy.

00:19:04:06 So that's why I come back to, it's a both and it's agent led spicy taste.

FRED VOGELSTEIN:

0:19:11.61  Go ahead, Elena, go ahead.

ELENA SUNSHINE:

00:19:12:01 Oh, sure. Well, I'll just take a stand. I think, it's consumer led. I mean, when you see the adoption of ChatGPT and just like people in my life, like the example you said that are using this every day. To me, as a manager of a team, I see my employees are using these tools. And like you mentioned before, there's a lot of shadow usage within companies.

00:19:38:22 And employees are demanding to have access to more powerful tools so that they can do their job better. And I think that comes from their personal experience, experiencing usually ChatGPT at home in their personal lives. So that's my hard stand.

APARNA CHENNAPRAGADA:

00:19:56:15 Not, not so spicy anymore now, but I will say, I want to unpack that a little bit, because I do think what's happening right now is, I mean, I saw this firsthand in search in the mid 2000s and early 2000s, where the learning curve was anybody who could type anything could get an answer. Now the learning curve is like, even for anybody who can speak to a human can now use a product.

00:20:22:05 So to me, it's obvious than that like that, this whole thing is going to be like individual led. So that's one. The interesting complementary force that's happening that I see in my day job in enterprise is that there is a lot of, like the cloud shift, for example, a whole bunch of, initial I wouldn't say skepticism, but hesitation.

00:20:42:05 ROI, all of those questions. There's a lot more of a different posture top down. That every CIO, every executive is like, how can we use AI? How can my employees use AI and so on? So, it's actually a very interesting end. And I don't think we've seen this kind of a thing. There's a bottom up, zero learning curve.

00:21:01:20 The whole world is like seeing how easy it is and how effective. And there's a top down thing. But they are not happening at the same time, at least from what I see, because the first wave of these models, I always say the model is the product in some of these cases. The first wave of the model, if you think about it, is amazing at answering questions because it's next token prediction.

00:21:25:16 That was the first wave. That's why you have chat bots. The second wave starting this year has been the reasoning models. For any of the agent stuff that Anthony is talking about in process automation. You're not going to be okay with a ‘Oh, it's sometimes it works in the enterprise’. So, you do need the reasoning model unlock.

00:21:45:10 And the third wave is going to be this multimodal, autonomous long-running thing. And once the models get better at that then we will see a lot more in the enterprise.

FRED VOGELSTEIN:

00:21:55:14 Diya, does all this like jive with you or is this like or is the world that you live in seeing this entirely differently?

DIYA JOLLY:

00:22:02:13 No, I think this does jive with me. I will start, it is definitely consumer led and that is I think we're very aligned. And I'll give you an example of why I think this is consumer led, a hilarious example. And this happened three days ago.

00:22:17:02 My son is a teenager and I hear him talking to a friend on the phone and they're talking about how that friend should approach a girl and his friend is talking to ChatGPT, telling ChatGPT stuff about the girl and asking advice on what the girl would like and how he should talk to the girl.

00:22:33:19 So it is definitely, definitely consumer led. And teenagers, I have no clue. Because the second my son found out, because he was on speaker and I could hear, and the second he found out I was listening, like, obviously he went away. And I found that fascinating. I think there is a difference. So, I think here's the difference.

00:22:50:04 I think one is what Aparna said the ability of technology to get into the hands of people, as just like ChatGPT has gotten, it's never been seen before. So that's one thing. I think the second thing is we are in a generation where this generation has lived through multiple technology changes.

00:23:11:10 If you thought about it like the technology changes are just getting more and more rapid and we're probably one of the first few generations that have lived through an entire cycle in our work life and our careers through technology changes. So we are pattern matching and going, this is big enough. Mobile was big enough, internet was big enough,

00:23:27:10 this is big enough that something's happening. Which is why I think this time enterprises are reacting even more than they've ever in the past to go. How do we utilize this new change?

FRED VOGELSTEIN:

00:23:36:08  Right. It sounds like you wanted to say something...

ELENA SUNSHINE:

00:23:38:03 I'll just add one more thing. Based off of what you said, I think, like, the technology itself is super accessible, on its face, but I think to actually get real power from it in a work context, again, like I said before, I think people are just figuring this out now, and some people are better at it than others.

00:24:01:09 And there's maybe not so much of an access gap, but there is a skill gap between where certain folks are and where other people are right now. And like you said before, your job will be replaced by the person with AI. So how do we make sure that we ourselves are the people that are good at using this technology and can be more productive?

00:24:27:13 How do we train our children to do that? And on the same topic, I think, I'm also concerned about the concept that we will automate entry- level jobs with this technology and we create a gap in our talent pipeline that becomes the future leaders of tomorrow. So, those, I think are all problems that are remaining to be solved.

FRED VOGELSTEIN:

00:24:53:00 Right. I think last question before we let you guys jump in. Let's talk about how this changes management, right? I mean, in a world where your employees now not only have like tools like ChatGPT, but also like vibe coding tools that, like, would allow them to kind of actually put their own apps together.

00:25:21:19 How does this change the whole top-down management structure that we've all kind of grown up in and gotten used to?

ANTHONY ABBATIELLO:

0:25:30.31 So let me give you three point of views from a research and practical view. Number one, it is changing the way leaders lead. And when I say leaders, I really mean executive leaders. The top of the House, how they lead, the way we used to introduce change into an enterprise, big or small, we would decide we're going to buy some technology or make some transformation, make some operating model change.

00:25:55:07 We'd figure it out. Then we'd start to roll it out and get people excited or involved in it at the bottom. That's over, right? It is changed. We're meeting in the middle. And so, leaders have to think about. First part is how do I activate the citizens within the organization and how do I drive that and enable that democratization of technology so that innovation can start at the bottom.

00:26:17:14 The second is, we always talk about humans at the helm. There isn't a process or a workflow that doesn't have humans, not just that are in the loop that are actually leading. It's humans that are leading digital and human teammates. And how that all comes together and how you change the way you think about leadership and driving different types of capability, both at a strategic level and an execution level.

00:26:44:08 And the third is, it's all about the skill and behaviors and culture. So, especially leaders of a certain age, if you are afraid that the technology is coming to destroy your job versus the technology is coming to enable or help us be more innovative, get to market faster, or be more profitable? You're going to constantly put up a guardrail or a blocker around that.

00:27:11:00 You need to, back to the first point. You need to enable the organization to be able to do that one. And then for yourself, have the skills to do it. I coach a lot of CEOs who are in this stage, driving this. And they'll admit vulnerability and fear around the change ultimately drives them pushing back and getting them to try it.

00:27:30:04 Share with their teams. Creating a positive AI culture and environment will ultimately make them not just be a better leader, but make the organization be more profitable.

FRED VOGELSTEIN:

0:27:40.61 Go ahead, Diya.

DIYA JOLLY:

00:27:39:22 I think, I agree from a leadership perspective, I think for many people working in an organization that are not at the leadership level, but are like in mid management, etc., I think creativity is becoming more important than was the process followed, right. I think problem solving on issues that were deferred to people higher up above them is flowing down more and more, right?

00:28:03:19 So for example, things that would be passed off to an organization like, hey, how do we design something now in PM? Like the PMs are being asked to design with a replete or something, right? Like can you prototype it and design it. And then I think that job functions are merging, right?

00:28:25:22 Like job functions are beginning to merge because now different functions can do more. So, I think what managers at the middle level are being stressed to do is even if they came up through a particular function, they're being asked to actually go across functions where they may or may not necessarily have the skills. And that, I think is the scary part.

00:28:44:12 The exciting part of it is you now have the technology that can probably help you do it. If you can ask the technology how you should talk to a girl, you can probably ask it how you should do another function.

FRED VOGELSTEIN:

00:28:53:14 Right. Well, Diya as you were talking about, that, one of the things that was going through my head was, there's an entire, I mean, in major tech companies all around the world, there are product managers and then there are also like people who are like programming, like in the trenches. And what if those two jobs merge and how many layoffs exist as a result of that? 

DIYA JOLLY:

00:29:19:06 Yeah. And I'll give you another example, right. Like for us in our customer success organization, normally the job of customer success is customer support. You pick up the phone and you try to answer, you get an inbound call and you try to answer a call. Now the problem has become easier because of AI.

00:29:35:18 Their job is flipped. So, can you get them to use more of the rest of the product, right? So, your job kind of merges well, it's almost semi sales, semi customer success.

ELENA SUNSHINE:

00:29:45:06 Sure. So I totally agree with everything you said. I'm a product manager so I am, I think as product managers we may feel very comfortable wearing many hats. And so, for me and for my team and my employees, this is really exciting because, now, like we said before,

00:30:07:14 roles are blending and the same person can come up with an idea, prototype it, create really high-fidelity, beautiful mockups, create the beautiful marketing materials, create a short video that explains their idea.

00:30:21:09 And, it just allows kind of the cream to rise to the top. Again, this is the concern about the skill gap, but it allows people who have great ideas to communicate them more effectively. And I think that's great. I think another piece, bring me back to the original question about management. I remember when OpenAI announced their research agent and it was $120,000 a year for this deep research agent, like that was PhD level.

00:30:52:22 And they were first kind of announcing this pricing. And as a manager, I thought to myself, oh, that's an employee. That employee could be running 24 x 7. How would I think of ideas to keep that employee productive all the time? And I think we could see teams that are augmented by an AI agent and think of them as an employee.

FRED VOGELSTEIN:

00:31:18:09  Well, I mean, in their company. What's the company that's getting a lot of retraction, open evidence that, is now making I think it's ad- supported now. So instead of, like, costing $120,000 to have access to, to be able to query like, and like as a doctor to have, like somebody to ask reliable questions to, the idea is it might actually be relatively free.  Anyway, go on.

APARNA CHENNAPRAGADA:

00:31:53:22 I think this is going to be when you think about management and organization, I think this is going to be the biggest shift since, I would say, assembly-line industrial revolution to, maybe the when the PCs came on board, this is the 3.0 of it, and I think it will hit us all at multiple levels. So at the individual level, as we all operate, I think it's true.

00:32:15:12 Previously I think intelligence and skills were the gatekeeping, right.  Now, every one of us has the super-intelligent thing but you call it AGI. That's actually beside the point. But you have a really, really smart companion copilot in your pocket. And so that's no longer the gatekeeping thing. It is going to be ambition. It's going to be who takes the flying car to the grocery store versus actually to Mars, right?

00:32:40:04 In terms of like, how are you being high ambition in using AI? It is going to be about taste and judgment, because you have a team of agents and it is going to be about every employee is now a manager, is a boss of these agents, which means you'll have to figure out how to steer, how to coach, how to direct.

00:32:59:16 These are all skills that typically people get much more further in their career. And so that I think is going to be an individual-level challenge. The next level of challenge and opportunities at the team level, by the way, we're seeing that already where we added meeting agents and these project-management agents into our teams’ product, and it's fascinating to see the dynamic.

00:33:23:04 It is what you said, Anthony, about multiple humans and multiple agents at work. And I have a team that's kind of like, it's very Meta. It's actually building a feature for that project management agent, using that project management agent. And one nugget I would say of insight there is that today we think about our wiring default is not asynchronous in 24x7 to your point.

00:33:45:16 So things like, hey, these things we are VSLMs, humans. And we have incomplete knowledge. We have incentive issues. So, turns out that the LLMs are 24 x 7 but also, they have like amazing — they are not token constrained like we are. And so, whichever team and we have multiple teams using it.

00:34:10:04 Whichever team is able to put those agents to work in a much more effective way, they're shipping in the order of weeks versus months and years. That’s at the team level. The organizational one is the one that makes me most ponder a lot. In fact, most recently I wrote about this, saying most of what we do in an organization is translation and routing.

00:34:37:03 If you look at an organization, there's a whole bunch of folks who are doing things, right. And of course, swim lanes are merging there. But they're all builders, whether they're engineers or designers. And then there's a whole bunch of information routing and translation that happens. And then you have an executive that's kind of making decisions.

00:34:57:18 If LLMs are really good universal translators, I think one of the questions with manager in their title in all of us needs to think about is, am I doing translation or am I doing transformation? And adding kind of the value and how am I being effective in helping decision making and judgment versus schlepping information one place to another because the LLMs will do it better.

FRED VOGELSTEIN:

00:35:20:06  Anybody who wants to ask a question should start coming up.

AUDIENCE SPEAKER:

00:35:25:22 You brought up something that worried you about the talent gap at the entry level. And yesterday we heard May talk about this flattening of organizations and working laterally. What are you hearing, clients your own companies, are they talking about this? What are they doing about this? Like, I know it's not solved, but what is the dialog?

ANTHONY ABBATIELLO:

00:35:51:20 I'm really glad you asked that question. It was actually the one…

FRED VOGELSTEIN:

00:35:56:03 The next question I was going to ask is what is the entry level? What does the entry-level job look like?

APARNA CHENNAPRAGADA:

00:36:01:21 I mean, this is actually what we are living through both in terms two dimensions to this problem. That one is how do you actually train folks coming in, like you don't want the first rung of the ladder if you will, right? If you have folks coming into the organization, how do you equip them with skills to use AI, right, to be able to get those skills?

00:36:20:05 I was talking to a law firm, like a partner, and he said, oh yeah, like I actually use a lot of AI tools to supplement the analysts and associates. And I said, oh, how do you have the judgment to kind of like, steer these tools? Because he's like, oh, I was an analyst, an associate.

00:36:38:14 I'm like, do you see what's happening? You need to kind of figure that out, right? So, I think one of the things you're talking about is what is the externship, apprenticeship, training model for these, like beginning rungs of thinking with AI, thinking with the machine and working with the AI and not just using AI as a tool.

ANTHONY ABBATIELLO:

00:36:59:06 Yeah. So, two things. One, the entry-level job in many industries is changing, especially when (Unintel phrase _____37.08) PwC, I mean, 25% of our business is audit, right. And so, it needs humans that read financial statements and then ultimately, create an opinion that's going to change, right?

00:37:20:07 We're going to see that be more automated and autonomous. And so, what we need is our associates that come in from undergrad that actually have AI skills that know how to work with those agents. That then we in a regulated industry still need a human to render an opinion. So, the job is changing.

00:37:39:16 Now the numbers is a different question, right? So, the job is changing. The interesting piece is we also have less humans. So, we're not going to be producing as many, particularly in the United States, that the number of people that are starting into these careers. And so, there is this how do we push people like other countries who are more mature and older than us into other types of careers and understanding of that, which then…

00:38:03:13 So that's why I was saying it's a social experiment because it affects how expectations we set of children like it's okay to be a welder, right? Welders in the United States make a lot of money and no one tells their kids, oh, I want you to grow up and be a welder. But they say, like, oh, I want you to be an engineer.

00:38:18:21 Everyone needs to be an engineer. I don't have kids. So, I hear people, my friends say that.

APARNA CHENNAPRAGADA:

00:38:22:21 But if you want to be an engineer, you should be in…

ANTHONY ABBATIELLO:

00:38:24:23 No, I'm not saying, but there's plenty of people who can't be an engineer. And then we have all this dialog between unemployed and job loss. And actually, there's not job losses so much gain. It's just not in the way people think of it today.

AUDIENCE SPEAKER:

00:38:38:21 A quick follow on question is just to throw out there, it kind of seems like we just need to really look at hiring practices and think about that pretty deeply.

APARNA CHENNAPRAGADA:

00:38:49:10 We should follow up.

FRED VOGELSTEIN:

00:38:51:03  All right. Next question. 

AUDIENCE SPEAKER:

0:38:56.81 Yeah, I have a question regarding the C-suite and what kind of skills should CEOs have in order to navigate the next decade. 

APARNA CHENNAPRAGADA:

00:40:17:19  I'm on the board of eBay, and I was on the board of another public company. And obviously we talked to a lot of companies and CXOs. I mean, it's hard to give a short answer, but I would give like two observations that I've been kind of thinking about.

00:40:38:01 I think one is this idea of like defining the business and the industry that you're in. For example, ages ago, I think there was this question about a bus and transportation companies, railroads and how that whole thing happened because they didn't think about themselves as in the transportation business right? Versus like, exactly and I think the question is today, for example, whenever there is a technology and platform shift, what I say is there is going to be a collapse of categories.

00:41:06:05 When the internet came about, before that, you didn't remotely think about a phone book and a local newspaper in the same bucket but once the search engine came, there both turns out to be the blue links on the web. So just like that, there will be a category collapse of things that are very different. Already, I see in my business — productivity, creation, search information, it's all blurring together.

00:41:33:04 So that's the number-one thing. Are you in a category that's going to collapse with the platform shift? If so, really think hard about how you want to play. The second one is more psychology I would say is which is the pasture. I think it is very tempting for incumbents to kind of like not lean in and avoid and think it's going away.

00:41:51:00 Diya is right this time. Many of the incumbents today were the upstarts in the internet shift. So, we are familiar that this is going to be big. That's knowing that intellectually but really leaning in and having AI facility. So, I even say go and actually be in, I see and like build things because you don't get intuition and learn about what the models can do unless you do that.

ELENA SUNSHINE:

00:42:16:16  I'll just add one thing if you want to jump in also. I was just having a conversation with someone who is an executive at a cable company and we were talking about the advent of streaming and how it destroyed the industry. And, yeah, I was reminded again, going back to the point that I made before that, it's not about what are the AI problems that my business can solve.

00:42:40:05 It's about like focusing on the core value that your business is driving. Can this product be faster, cheaper, better, more delightful for my customers? Can I provide more value? And so I think there's again, when you're in this kind of incumbent spot, there's, can be a failure of imagination about what could disrupt what the business that you're working in.

00:43:03:15 And so I think, my advice here is just to focus on delighting your customers because someone's going to take your bag if you don't.

DIVA JOLLY:

00:43:13:04 I think, the CEO really needs to understand a couple of things and then have the courage to go through the change to make those things happen. So, the first is what is truly, deeply your company's value proposition and skill sets? And what I mean by that is a lot of companies confuse their product or their customers with their skill set.

00:43:39:19 If I look at Xero, I'm going to take an example to make this concrete. We provide a piece of software that does your accounting taxes, your payments and your payroll. But a real value proposition that distinguishes us, is in the way we connect a small business with the advisors they need. And our entire product is architected around that.

00:44:01:08 And that is what makes us successful. Our entire go-to- market motion is architected around that, and we are unique in that worldwide. So, if we look at it from that lens and go, that is the value prop we're providing, that's not going anywhere. Connecting a small business to an advisor that can help and an advisor to a small business where they can help etc. that's not going anywhere.

00:44:20:19 How do we use this technological shift? We give up the entire baggage we have. This is what our product looks like. This is how we do business. Is there a business model? And you need a CEO that can go, okay, given this technological shift, how can I make this value proposition more deep uniquely? I think that's what's going to stand out.

00:44:39:02 And then obviously there's a ton of courage. Do not use a bad word. There's a ton of courage. You have to have to then change your entire business in that direction.

FRED VOGELSTEIN:

00:44:51:23  All right, let's go to one more question.

UNKNOWN SPEAKER:

00:44:55:00  So thinking for the leaders or managers, so it's kind of building off the CEO question, when you're in an industry or a business that has to get stuff done, like we have to continue the business that we have and sustaining the business that we have, but we also want our teams to be creative. Like we've heard this throughout the last couple of days, AI is going to liberate you from these mundane tasks, and you're going to have all this creativity and time.

00:45:21:10  Okay, cool. Except when am I going to do that? And I have these goals that I have to hit or my team has to hit or whatever for the quarterly whatever it's OKR or goal or whatever the metric is. So, I'm curious to know what you've seen or done as leaders for your teams.

00:45:43:10 Like what metric are you asking people to hit to let you determine if there is the kind of creativity that you want or creative thinking, which is sort of this hard-to-measure reality that I think it has to be there for people to do this kind of AI exploration we're asking them to do. How do we do that?

00:46:00:07 Like, how do we structure the goals of our team in a way that lets people do that? I don't know. I'm just curious how you've done it. How you’ve seen other people do it?

FRED VOGELSTEIN:

0:46:12.51 How do you keep people from using AI to cheat? I can’t get a better bunch.

ANTHONY ABBATIELLO:

00:46:15:02 So, that's a good point because I think what you have to do is you have to you have to liberate them and make them feel like this is actually enabling their success, like their productivity.

00:46:24:10  And you have to think about ways in which you're going to incent them. But the part that I think you're really getting at is about creativity, and not everyone is created equal. So, the key piece that's still a gap for many organizations, the ones that are getting it right are giving that skill

00:46:40:20 and development to their staff, to their people that say, you need to get the skill on how to use AI, but also the human skills are more important for the humans that are still in these jobs. And so what does that mean to be creative? What do I do with the time that is that I came back because AI is enabling me to have greater efficiency.

00:47:00:05 How do I use that time? How do I structure it? And then you as the leader have to make that positive culture in that environment that allows that team to feel like, hey, it's okay if I spend two hours thinking about new ways or new innovations and bringing those to staff meetings and ways in which that they're going to get rewarded for that versus, oh, you didn't hit your number, but this is a great idea. You have to create that culture, and you have to stand by it and live by it after you give them the capability.

UNKNOWN SPEAKER:

00:47:23:02  Is that like creativity, OKR, I'm like practically like, what does that look like to do that?

ANTHONY ABBATIELLO:

00:47:29:13  All of it should be in service of key results of your organization like one shouldn't need an OKR for what we're talking about. It's ultimately in service of the business result. And if you're doing something that's not in service of you shouldn't be doing it.

00:47:46:05 But if automation for autonomy, if agentic AI is going to deliver that greater effectiveness of the business, then using that creativity for you is going to be something else that's different than for me.

FRED VOGELSTEIN:

0:48:00.51 Each of you has like time. We have time for like two sentence answers here.

DIYA JOLLY:

00:48:02:08  I think, I agree. The thing I would say is innovation or creativity often happens at the bottom where people have time and not at the mid management or senior management level. So, one of the things we've found a lot of success in is like we will hear of things people are doing, new employees coming in, new engineers coming in, engineers at the lower level.

00:48:25:07  You just have to make sure everybody hears of it, because once people hear of it, then they will start adopting it. So, you go to these pockets of creativity and that's what has worked for us.

APARNA CHENNAPRAGADA:

0:48:38.21 I think the mimicry. I'd say we've run, like my teams done boot camps where we say take one week or two weeks and just like you get all the tokens you want and all the tools that you want, go build stuff.  And initially there's no pressure about kind of like making it add to the bottom line, because what I found at least is that that kind of clamps your thinking.

00:48:59:14 You just kind of like, go for it. But not everybody is going to sign up. It's a concentric circle model. You want the folks who are already AI build and then kind of like get to the folks and mimicry.  And then, the diffusion happens as it does with any change.

FRED VOGELSTEIN:

00:49:09:11  All right. That's all the time we have, give our panel a hand.  Thank you all.