← Digital Disruption with Geoff Nielson
Digital Disruption with Geoff Nielson
Jobs After AI: Futurist Ian Beacraft on What Happens When AI Does All the Work
01:03:43
Loading summary
Transcript52 lines
- [00:02]
A
What happens in these next two years is all that will matter when people look back and take a look at what you did the last 30. I'm sorry, but they don't count.
- [00:14]
B
Hey, everyone. Today I'm super excited to sit down again with Ian Beecraft. He's the founder and chief futurist at Signal and Cipher and an undisputed thought leader at the intersection of AI and work. I was a huge fan of our last conversation and super impressed by the breadth and depth of insights he brought to the table. He's on record as saying he thinks the next two years are going to be more impactful than the last 30. So I'm really excited to get his take on where work is headed and how helpful the latest AI tech really is. It should be an amazing conversation. Let's jump in. So, Ian, it's been, you know, I think close to a year since we last talked, and I'm curious, just, you know, reflecting on, you know, all the changes in that time and, you know, it feels like dog years these days with, with AI and the pace of technology does. What do you see as kind of the. The biggest trends or, you know, the biggest things on your radar, both in terms of what's changed in technology and in our posture toward technology?
- [01:20]
A
I mean, it could take conferences to kind of catch up on just what's happened since then, but I think one of the things that inspires me the most is the conversation has switched from, hey, this stuff is cool. Look at all these things these tools can do to, oh, this is challenging our whole concept of what an organization should look like, what work should look like. Those who've been leaning in and really spending the time getting to know the tools and how they can implement into the work that we do are moving beyond just how do I become an advanced user of the tools to how do I reshape the actual notion of work, my own impact, my relationship with these things. And it's become a question of identity, organizational design. There are much bigger questions than just technology now, which is fascinates me.
- [02:10]
B
So along those lines, one of the narratives you've challenged is just this talk of jobs and how AI is changing jobs and tasks and more of that kind of design layer. Can you, can you unpack that a little bit? What, what's, you know, the wrong narrative here and what's, you know, a more productive way of thinking about it, for sure.
- [02:29]
A
So the, the narrative, I think that is too small is that this is about using tools. AI isn't a tool. Tools are A new way or more effective or efficient way to do a task. I think we can say I'll see safely that AI is much bigger than that. It's a platform that, that rewrites the rules of entire economies and ecosystems. So if we think of it as a tool, we ask tool like questions, what can I do with it? What does it do for me? Those are very small and narrow questions. When we start looking at how it rewires the rules of economics, how it rewires the rules work, we start asking much bigger questions. So it changes our relationship with it of how do I use it to, how does it help me reframe what work even means for me? And that's the thing that I think that people have been grappling with the most, have kind of gotten past this. Oh my gosh, it threatens my work to. It reorients my posture towards work significantly. And that's going to be an identity shift for a lot of people. That's why there's so much anxiety around this is we're looking at automations, but it's also starting to challenge our paradigms on what work is, how organizations are designed, and how technology plays a role in that. So I mentioned organizational design. And the thing that I've been really leaning into lately and investigating and studying is the way that we design. Our organizations today are a result of the physics of human work. They weren't designed for AI and agents collaborating with us. And as a result they're challenging literally every institution we can think of when it comes to the functions, the variables and the structures of the work that we are so familiar with.
- [04:10]
B
One of the insights I remember from our last conversation is that you told me that the threat that people and organizations are facing right now is not necessarily AI itself. The greater threat is having these outdated mindsets like you just described about what an organization is, how it should function. Do you still believe that? And you know what, if anything, have you seen in terms of mindset shift in the past year?
- [04:39]
A
I actually believe it even more so now than I did back then. And what we're seeing bear out is organizations that lean so too far into the technology and say, hey, we can one for one replace an individual with a set of agents because it's cheaper, it's faster, it's more effective, it can answer customer questions faster. They're learning. It's not that simple. What they're learning is that there's two elements to an organization that they have to kind of wade through, and that's coordination, which is what agents do very, very well and what most people think is the only thing work is all about and culture. And oftentimes we find that culture is actually doing the work of coordination, is kind of doing a couple things all at once. And we just can't separate the two. We don't know how to identify what's what, what's coordination versus what's culture. So I'll give you an example. Your morning standup is both culture and coordination. You're imparting knowledge about what's happening, what the status is. But there's also cultural elements to that in terms of the vibe, the organization. There's teaching, there's camaraderie, there's a bunch of things that come with that. But sometimes the culture is actually doing the work of coordination where the coordination or the structure of the organization fails. How many times have you been in a standup or even any meeting at all where something's not going right, there's a miscommunication, there's friction of some sort where things just aren't working properly. And what happens is we step in as humans to fill that gap and we just figure it out. That's because the system that we built for that coordination had failed and we as humans with our culture has stepped in to fill that. So culture becomes like this load bearing structure in an organization and you start to see the, you know, how these two kind of pull apart at the seams where culture and coordination are necessary. But the reason I bring that up, and I'm so adamant about it, is because when we start talking about agentic AI and AI replacing roles, we're ignoring a huge part of that equation. We're just focused on the coordination and execution component of it.
- [06:43]
B
Well, there's an implication there, if I'm, if I'm reading that correctly, that you know, leaders, CEOs, boards who just take a machete or an axe to parts of their organization and say, you know what, we're just going to rip and replace this with AI, that there's going to be some pretty detrimental unintended consequences if they don't consider the broader cultural piece. Is that a fair reading?
- [07:09]
A
Yeah, and I think we've seen that kind of come to bear in the headlines too. There's the classic Klarna case where they let go about 700 contractors. I know that's been in dispute, but it seems to be pretty reliable telling of a story where they essentially let go of about 700 contractors and full time employees and let more of the agents handle Those requests. And what ended up happening was even the CEO, like the C suite, had to kind of jump in and get involved in some of the ticket resolutions because people were getting frustrated with the responses they were getting. It wasn't actually solving the challenge. They were getting informational responses, but they weren't getting solutions to their challenges. And that's a huge difference. The agents that they had put together had been really good at relaying that information, but not at getting through the points where they could actually get to resolution. Because you have to hop between systems, you have to be able to align with the person on the other end, really understand what they're actually looking for and connect all those things. That's something that agents aren't really good at yet. And as a result, they ended up hiring a lot of those roles back in order to fill those gaps. And it ended up costing them quite a bit of damage in brand and reputation and a lot of money to kind of re engineer that whole process. So that's just one cautionary tale. There's several others. But organizations that have leaned too far in that direction are now starting to see that there are multiple components to embracing AI. It's not just use the tool. You have to fundamentally rethink the workflows, the systems, the departments, the job descriptions, the functions, all of that when you implement AI. Because when you get efficiency in one place, you don't just eliminate the bottleneck, you shift where the bottleneck lies. And it's that second order impact that most companies have not grappled with.
- [09:00]
B
If you work in it, Infotech Research Group is a name you need to know no matter what your needs are. Infotech has you covered. AI strategy, covered. Disaster recovery, covered. Vendor negotiation, covered. Infotech supports you with the best practice research and a team of analysts standing by ready to help you tackle your toughest challenges. Check it out at the link below and don't forget to like and subscribe. So I want to talk in just a second about the right way to do this, but just before I do, I want to pick on another implication here because one of the themes that I'm reading a lot about and hearing a lot about is it feels like there's this sort of angst and dread at the employee level about this shift and about this technology, regardless of whether or not it's a tool. And the implication I wanted to ask you about is that it feels like in this world, and as soon as we're talking about rewriting systems and restructuring whole workflows, it feels to me like executives, like boards, the people at the top have a lot more impact here than maybe they've had in the past. For good or for bad. Right. Their ability to take a company or an organization in the right direction or the wrong direction feels like it's a much more dramatic, there's much more variance than maybe you've seen in the past. Do you buy that? And if so, what are the implications of that reality?
- [10:32]
A
Would you, you mean variance or like oversized impact in general?
- [10:37]
B
What I mean is variance and impact that, that the amount of good you can do and the amount of harm you can do are both feel like they're magnified.
- [10:45]
A
I would agree with that. I think he's absolutely right. Because at that point in time you're making macro decisions about the environment of the workplace. And what people don't understand, they're still treating this like it's an IT issue. And this is about provisioning licenses, it's about replacing like for like. It's about reduction in force. Because whenever you see something that is an efficiency based technology, the first thing that most boards are going to look at is say, ah, that is code for layoffs or that's code for reduction in force. Because I'm getting that efficiency, I should be able to kind of replace those two. And it's not that simple because it isn't just about efficiency and speed and scale. The thing is, when you put it in a play of an environment like this, when it's ubiquitous and everyone has access to efficiency, speed and scale, you get a race to the bottom. You can only cut yourself through efficiency so far. So when you lean into that direction a, it's a commodity. So everyone has access to the tool. It's not like this is some sort of super secret weapon. So whatever you do to just kind of replace like for like has no strategic advantage, no competitive advantage. And when that's the lever you pull, you're easily going to be out competed in some other way with by somebody who's actually thought through the problem a little bit harder. And realize that when you put this into your environment, you have to rethink the environment entirely itself.
- [12:15]
B
I, I completely buy that. And to your point, it feels like this is a daunting challenge. And again, it sounds like you were getting at the fact that the challenge is not going away. And maybe in some ways it's even a bigger challenge than it was a year ago. So how do we get this right? What is the right way to do this? And if we're going to have a Coffee and a chat. A year from now, what has to happen so that we can say, yeah, things have been going good versus we're losing even more ground? Yeah.
- [12:48]
A
So the, the thing that most organizations have not even thought about when it comes to AI is how do we actually get these things to understand and operate within the context of our company and understand the DNA of our company? Most are just kind of using the tools as they are and handing them to employees and say, go get more efficient with this. And when you don't have the standards of your department, your team, your role, the job, the function, the workflows embedded in the data sets that these tools are working with, you are defaulting to the standards that come with the tool itself. So your standards are the standards that OpenAI is setting for you. Your standards are the ones that Anthropic is setting for you by their defaults and their understanding of your company. Ultimately, what happens is, it's kind of like if you got a. This is back in the early 2000s, 2000s, and you started pushing all of your data to the cloud. Well, let's say you got a new cloud provider, but you had no data in it and somebody else put the data in there for you, you don't know what you're going to get. That's kind of what's happening with AI tools is when you don't have your data formatted and also set up in a way that these tools now understand the context in which they are operating in, they're not going to give you what you want. They're. They're not going to be reliable, they're going to hallucinate more frequently. They're not going to give you the outcomes that you're looking for because they don't understand what success looks like. They don't understand the failure conditions of the workflows that you're trying to embed them in. And that's all sorts of context that most companies have not spent the time to articulate. And I don't mean just hooking it up to Azure or to your OneDrive and saying, hey, go read the documents. I mean, really spending time to figure out how do you encode these things into the data sets in a way that is designed for these systems, not extracted from existing documentation by the systems, but really actually practically hand tailored to these types of things.
- [14:46]
B
Right. It's a way to kind of encode that cultural piece and all kind of the latent, you know, organizational wisdom that's sitting in everybody's head and how can we actually use that to parameterize, you know, any sort of agentic work that's happening? Is that fair?
- [15:04]
A
Yeah. And, and what you've hit on there too, is also one of the sources of anxiety for people too, because when you start leaning into that and say, okay, I'm going to help you encode the expertise that you have, the, the wisdom that you've gotten from being here for 15 years, the cultural context that you have, the implicit information that allows you to understand the shortcuts of the work that's done here, that becomes a question like, well, if I give that to you, does that mean you can replace me? And ultimately, what we found is when people do lean into that and they start to take what makes them who they are within the context of the organization, they end up getting augmented. Far more than anything that they were doing is getting replaced, because their impact actually gets felt further throughout the organization. So when I'm starting to encode my own expertise, one, I can use that my own myself, I can start to divert a lot of the administrative work that typically would have demanded my time and my expertise to the systems that I've built out for that. And also the inbound requests can start being handled by the agents that you've set up that now understand your expertise, your cultural context, it starts to actually return the benefits we were hoping for in the first place, while really elevating the individual who's been able to set that up.
- [16:22]
B
So it feels like I buy all of that. It feels like it's predicated on a certain level of trust though, right, between employee and employer, that it's not going to be, as soon as I finish penning this thing, as though you would be doing it with a pen, that they're going to say, okay, great, now that we know everything of value about you, you're out the door. And, and so, you know, if we forget for a second, you know, the, the actual AI side of things, what do employers need to do and what do they need to signal around the contract with employees? And I mean that in the loose sense of contract, not in the, yet, you know, the very specific sense of contract, social contract, more of the social contract. And like, it feels like a lot of the headlines we're reading, there's some very high profile breaches of that contract. So how do we get past that if we think this is going to be a necessary precondition for actually augmenting everybody and making this thing work?
- [17:24]
A
Yeah, I think that's probably one of the most important Questions of the moment. And we definitely hear examples of that in the news every other day. The thing that most leaders forget is that people look at what you do, not what you say. And if you are not modeling the behavior you want to see, there's no way you can tell them to go forth and do it. Leaders have to lead from the front, not from. Not from the rear. And I'm not seeing a lot of C suites do that. Most of them understand that AI is important. That's the future of the company. But they're usually saying, hey, go and do this thing. And that's not very inspiring. The other part of this is they. People need a clear vision of where they're going. They need a crystal clear articulation of what the future of their company looks like and described in such a way that they can see themselves in that future. Can I see myself in your articulation of the vision of this company that I work at? And can I see myself succeeding in that vision that you painted? And if I can't, that gap in trust is even wider. There's even more that you have to do in order to bring me forward. And that's what I think a lot of organizations are struggling with right now. They have not been able to articulate what their future looks like, and therefore, they have not painted a narrative people can buy into. And when that narrative's not strong, they start to look for the stories elsewhere. They look to the news for the narrative that they should be telling themselves, okay, if I do this, then that means I'm getting laid off. If I do this, that means I'm replaceable. Whereas if the narrative is strong and then people see other people in their organization succeeding by doing that, it changes things entirely. When people see others who look like them, who come from the same background as them, who work in the same environments as them, who are part of their culture succeeding, then they start to look around, say, okay, that can be me too. But when the only examples they have of people succeeding don't look like them, don't work in the same industry as them, don't feel relatable to them, then they're disconnected from that story of success. And the only narrative that they can tell themselves are the loudest signals coming from outside of that environment.
- [19:28]
B
I think that's a really interesting and compelling point. And it sounds like, Ian, your experience is the same as mine, which is just, we're short on vision, just across the board. Like, that's one of the bigger challenges that we're facing. And my experience working with a decent number of C suite across a decent number of industries on this, and I don't mean to dunk on them, but I'll take a deliberately, you know, provocative approach here, which is it's a lot harder to invent a vision from a blank sheet of paper than to, you know, pull it or, you know, kind of craft it from other sources. And it feels like there's a lot of leaders out there, you know, just looking around the room to their peers or to consultants, saying, can you. Can you tell me the vision? What's. What do you think the future of the industry is and waiting for someone to hand them the answer versus actually doing the hard work to come up with the answer. And so I'm curious, first of all, if you agree with that, but if you do, maybe if you can share some counterexamples or some positive stories about organizations you've worked with that have been able to kind of crack that code and actually kind of demonstrate the agency to change that, that struggle.
- [20:49]
A
Absolutely, yeah. I'll give both sides of that equation. I'll start with the one that I feel like is less inspiring. And practically three or four times a week, I'll speak with either a C suite executive or a board member who says some form of this raise of, hey, I've got two years left. I just want to ride this out and I'm done. And I say, I get it. All of your predecessors have had the opportunity to do that. Everybody that you've ever watched in this role has had that opportunity to say, hey, I've got two years. I'm going to show up, I'm going to do my thing. But I'm not, like, giving everything I've got to this because I've spent the last 40 years doing exactly that. I'm not going to coast, but I'm not going to give it all. And my response is, here's the deal. What you do in these next two years, if you decide to lean out, that will be your legacy. If you decide to lean in, that will be your legacy. What happens in these next two years is all that will matter. When people look back and take a look at what you did, if the last 30, I'm sorry, but they don't count. If you decide to lean out because this moment is so important in terms of what you do, the momentum you build now accelerates by virtue of the way that the technology is changing and the way that it's affecting every sector of society, every sector of business, every part of the organization. This is not incremental, this is exponential. So small motions, micro actions have macro effect, macro actions get exponentially bigger. So it's not a choice. At this point in time, you don't have that choice, really. So that's one end of the coin. The other is organizations and leaders that have really understood this is a pivotal moment, not just for them or their organization, but for work as a whole. They're starting to toy with the idea of not how do I make my company more efficient? How do I fundamentally reimagine what our company could be? Or do I start an entirely new function from the scratch that can then inform the rest of my company? And those are much more visionary approaches. You don't want to upset the apple cart, completely reverse your business model overnight. But we do need things that can show us a radical approach to how work can look by using the native functionalities of these tools. That's not something I'm seeing a lot right now. I'll give you an example What I mean, most people look at this as a way to augment existing workflows and just make things faster. That's an incremental change. What they're not doing is realizing that this is something that has to force you to relook at not just a function or a task or a workflow, but the values of the organization. What are the things that we're actually. Where is the value of the work that we're doing? Is value moving in our industry? Is it being captured appropriately? Because if we're not paying attention to where value is shifting, we might be getting really good at the wrong thing. And that's just as bad as not paying attention in the first place.
- [24:02]
B
Again, a lot to unpack there. And I really like that approach. So it feels like what you're saying, Ian, is that if you're going to do this successfully, it probably starts with zooming out and actually reimagining your sector or your industry and thinking about where does value live and how can we be part of that value ecosystem?
- [24:22]
A
Is it.
- [24:22]
B
Is that a fair statement?
- [24:25]
A
It comes from both directions, actually. It comes from having a much more macro perspective and also a much more micro perspective. So I'll start with the micro spectrum, because this is something that. That's not, I don't think, intuitive. But the companies that have been able to articulate the implicit knowledge of the organization, kind of like what we call the soul of the organization, and been able to encode that in a structural identity, have been really, really effective. So an example would that be how you articulate your values, your mission, your vision, all that kind of stuff? Most of that's written somewhere, but they're usually just like hollow platitudes for a branding exercise. How do you actually turn that into what it actually means in the organization and put that into a document that agents can use? Not just values, but once you've really articulated your values, we go from the why do we do what we do? And I mean at a very robust level to how do we do that and what is it that we're doing? So why do we do it, what is it that we're doing and how? On top of that, that is not just your SOPs and your workflow documents, your, your blueprints. It's a text based articulation of exactly what those things are. And that's actually something we spent two and a half years doing for ourselves at Signal and Cipher. And the results have been profound. The fact that we spent that amount of time doing that, and we made that bet that early, has been a huge wind in our sails in terms of being able to take advantage of how agents behave. Because what we found is that if you articulate the things that we take for granted, the implicit knowledge that we have across the organization and the structure of the organization, that never gets written down. And you actually put that in a heuristic data set and start with markdown files. Doesn't have to be anything complicated. This becomes the structural fallback for agents. So even if you haven't given them great instructions, they still have a really rich and robust understanding of the context of your organization. Why do you do what you do? Okay, this is the why. If I'm being instructed to go do this, I now have background knowledge as far as why I'm doing that and how I should be functioning. Even if I run into challenges that my, my set of instructions don't give me to solve that for. So that is essentially a set of operational and governance procedures that are now embedded in the system. So what that does is takes this thing that we typically put on the outside of the apparatus of the organization. Let's talk about governance and compliance. Two things very few people think are sexy. And if you take that and encode it into the foundation of what your agents are working with now, it's not just this thing that sits outside the organization. That is occasionally applied when compliance is in effect or when you have to evaluate something. It is fundamental to the action of the agents and it flips the script from compliance being this thing that slows things down, or governance being this thing that interferes with progress, but to the very thing that accelerates progress. So by being fundamental from the bottom up, being almost this atomic unit of agentic work, what it does is you're able to accelerate the pace of adoption and production, because literally everything these agents do is in compliance. It is in keeping with the governance policies of your organizations. It works with all the operational procedures that you've already declared and constructed. So you're never going to get something from the agents that is going to be putting you afoul of regulation, governance or operational procedures. That's a fundamental reframe of the way almost everyone I'm seeing is approaching agents.
- [28:05]
B
It's really interesting and I want to talk about that reframe and how you actually put it into practice. And as you said, you know, governance and compliance are, you know, some of the least sexy topics out there in technology. And if, you know, if we don't want anyone to watch this video, we can label it something about governance.
- [28:21]
A
We can do a whole upset on that.
- [28:22]
B
Yeah, exactly. And that'll ensure we get exactly zero views. I've made that mistake before, but that's another story. So governance and compliance, these are areas, as you said, that they're often either buried in the organization or they're in some basement or they're looked down upon. Maybe they're in it, maybe they're in a compliance or risk function. If we're going to flip the script and we're going to change their prominence and their value, who is doing this work? Are you taking those existing teams and saying, hey, IT governance, IT risk team, encode this into the AI. Are you spinning up a new team? How do you realize this kind of reimagination?
- [29:02]
A
Absolutely. So this is the kind of work that will be happening in every sector of the organization. So governance and compliance is just one function of it, but this is something every individual employee will have to do for their own work to some extent. I'll give an example. Most people have realized that AI writing sucks. It doesn't sound like them. It uses EM dashes everywhere you go. It always has those phrases we've gotten rid of, delve, but now we have. That's the unlock. Or it's not this, it's that they're like signs that everyone can tell that AI wrote this. However, there are some writers who've spent a lot of time deciding what they don't know, want in their writing that AI puts in the writing, and then they've also spent time saying what makes my voice uniquely mine? That when I write it, you can tell it's me. That's hard to do because that requires that you take time to articulate something you've never really had to explain to somebody. There's a great cartoon from. It's Snoopy. It's Snoopy. And I'm forgetting the bird's name. But he asked like, how do you fly? It's like, I don't know, I just do it. It's like, well, show me. And all of a sudden when thinks he thinks about it, he falls to the ground. It's the same thing for most people. When we try to articulate things that we do naturally, we have a really hard time to do it. The problem is that's going to be 90% of the work you need to do to encode the knowledge you have into data sets. But those who've done it have 100xed their output, their capacity and their capability. I first watched my co founder do it two years ago. I've had the opportunity to do it for myself and the difference is profound. The challenge though is when you start to do that, you start to articulate your voice, your values, your function. It does feel like you're giving up a piece of yourself and saying, okay, now the machines have that, what's uniquely mine. The interesting part is that this unlocks all sorts of opportunity for you because you actually being amplified and your impact is being scaled by using agents to now affect your vision rather than the work itself. So what happens is we're now creating this hierarchy of different modes of work. So for the last 150 years, we have been responsible for execution. It's the artifacts we produce, it's the projects we complete. We as individuals are responsible for like the actual, the act of work. And that's some. That's the very thing that agents are good at. Now, just because they're good at that doesn't mean that we're out of work. Because what happens is we now have to elevate ourselves to a different level of abstraction. Just like the Industrial revolution took us from physical labor to mental labor being our value, we're going through another level of abstraction now where it used to be that we were responsible for doing the work. Artifacts, projects, outputs. And now we're moving up to designing the work. So that's workflows, that's automations. That's essentially like looking at systems and how those systems come together. So we're looking at a series of those things. That might have been dozens of outputs and tasks kind of thrown together, but now that we're looking at them from a higher elevation, we're also looking at the surrounding elements that change the impact of those things, because we're seeing elements and signals of change internally and externally. So when you're starting to look at design of work, you have to be very cognizant of how the external environment is changing around you. So as AI models change, as geopolitical events become more important, as the competition starts to move differently, as value starts to shift, those are external signals that things are changing. Which means you, as a designer of the work, needs to shift what works looks like to capture where value is going. That is something that does not stop. The world doesn't stop. It's never static. It always continues to change. So that's where the value of most work is going to shift. You're never going to run out of challenges to solve. So the people who are starting to see work that way aren't finding less work to do. In fact, almost everyone I know that's leaned into agentic work and Claude code and agentic coding is losing sleep left and right, yet they're still being 100 times more effective, but they're finding that there's so many more challenges that they can solve that they can't stop themselves from going and solving those things. So there's not less work to do. Trust me, there's plenty.
- [33:31]
B
Right. And, I mean, it's an interesting sales pitch saying that if you use AI, you can't sleep. But that's, you know, if they can't sleep because they're excited, I guess that's a slightly different story. But there are. I really like that entire model of designing the work and rethinking it and actually taking the time to kind of step outside your own tasks and roles and responsibilities and rethinking, how could I do this better? How could AI assist me in doing this better? There's two scenarios that it feels like we're sort of talking about at once. One of them is any employee independently saying, I'm going to do this activity and I'm going to rethink what I can do in my contributions to this organization. And the other one is hearing from your boss or your boss's boss or your boss's boss's boss, like, thou shalt. Like, it's a demand that everyone in the organization do this. We expect you to do that. And the reason I bring that up is I feel like there's a Very different flavor to both of those. A very different feeling of agency as an individual. And so, I mean, what's your guidance there? Is this something that people should be doing anyway? If you're a leader, how do you encourage it without being dictatorial? How do we, like, is there a way where both the individual and the organization can thrive?
- [34:51]
A
Absolutely. And to your point, 99% of organizations are getting this wrong right now, which is exactly why it does feel dictatorial, because we're in a moment where work is shifting so profoundly that you won't have a choice. It's like saying, I'm not going to learn to use a computer. There were people who said that when computers became a thing, There aren't too many holdouts anymore. That same transition is happening. It's just a much compressed speed. The challenge, though, is when you say, thou shalt people resist being told what to do in a way that they feel they have no power to influence it. But if you say, hey, the fundamental workings of our company are shifting as a result of this huge change in the world. It's not just like, hey, we need to use AI, it's the wrong approach. It's, the world is shifting. You feel it? I feel it. The organization is going to have to look differently in order to thrive in this new world where things are going. This is a challenge that I, as a leader, can't solve on my own, that no one person has a perfect answer for. This is something that we're going to have to solve for together. And I'm going to need your help as somebody who is an expert in the space that you operate in, to help me solve that. We need to come together and figure out what those solutions are. And it's going to take some time because there's no roadmap. Nobody's given us the blueprint. Nobody has the answer just yet. We're still figuring this out as we go along. And everybody in the world is in R and D right now. We're all thinking about what this means for our jobs, our roles, our employment, security, the organizations and the industries that we're in. So if we come together, we decide that this is what matters, we can start to figure out ways to go forward. But if all it is, is, hey, I don't know, go do it, people are going to find someplace else that has a much more collaborative way to do it.
- [36:42]
B
Right. And as you were. You were saying that. I was processing that. There's actually. It feels like there's sort of two electric rails here. There's that way of saying, you know, I'm forcing you to do this. And there's also the other way that, you know, I don't know if you're seeing happen where leaders just feel like they're going to, they will do it themselves, they're going to do it non collaboratively. And that doesn't feel like it's the right approach either. And so I like your kind of middle ground of like, look, we all agree this has to happen in some capacity. How can we do this together to everybody's benefit and hopefully have enough trust as an organization that people buy that and believe you versus actively resist?
- [37:21]
A
Absolutely. And I'm definitely seeing certain leaders saying, hey, I'm just going to do it myself. That is absolutely happening. Those who've leaned in and understand the power of the technology oftentimes say it's just easier for me to do it that way. It's kind of like when you're teaching somebody new. It's like if you understand it better than the other person, it's easier. It takes less time for you to do it yourself than to train the other person on how to do it effectively. But you scale significantly better when you actually have the patience to do it right?
- [37:50]
B
Absolutely. So maybe with that, Ian, I'd love to hear a little bit more. Just practically you mentioned that this is something you've been working on with Signal and Cipher for a couple of years now. Just share a little bit more about your insights. What are the things you've gotten right or wrong and some of the biggest takeaways that you try and share with people around that.
- [38:09]
A
Absolutely. So the right and wrong is an interesting track to go around because we are a fan of the, the scientific method of have a hypothesis and test it out and test it out at scale. Because we're in a world now where execution's cheap, failure is cheap, and it's fast. I talked earlier about the fact that all the fundamental assumptions of what makes an organization work are kind of falling by the wayside. And one of the biggest ones is that we need to spend a lot of time thinking and planning before we do so. Our organizations are kind of built around the assumption that execution is expensive and mistakes in execution are far more expensive from a cost perspective. Reputational damage, you name it. And that world required that. We spent a lot of time researching, doing synthesis, analysis, competitive analysis, you name it. This is the world that consultants live and breathe in all the time, where you can kind of do the planning and then you hand it off for Execution internally in our organizations, we do the same thing. We're figuring out what the markets are doing, we do the market planning, et cetera, before we do any of the execution. The way we look at it is we now say that planning is rehearsing your assumptions. Building gives you proof. And because execution is so cheap, it's now cheaper to build the prototype than to even have the meeting about planning the prototype. Just think about that for a second. It would actually be cheaper for you to spend three hours coming up with a prototype that shows people what the idea is about, that expresses the concept that you're trying to get across, rather than saying, hey, we're going to get all the people who are responsible for this. And then we're going to go through the racy and we're going to talk about responsibilities and timelines and we're going to talk about the future meetings and, and then we're all going to go off and do our planning and everybody come back and say, what'd you think? And like this. Six weeks later, you just started actually thinking about the project when somebody said, hey, I'm going to spend three hours banging this out. Come up with three different prototypes. We'll deploy one of them internally just so we can get some responses. And you walk in two days later with data like actual hard experience that shows what's working and what's not. That's the world we live in now. And the idea that we should still be doing this very slow and measured, almost defeated defensive posture means we're not leaning in and understanding what the native capabilities of this technology are. We're just sprinkling AI pixie dust on top of the organization saying, hey, we are now AI enabled. So that's one of the biggest things that we've seen is when you end up just building your, the fidelity of your data and your signal is so much stronger than your opinions. So you get to simulate and get that done. The other is that we're starting to see that all of the assumptions of the way an organization functions I said earlier, is kind of a, an artifact of the physics of human work. Things like job descriptions, departments sign offs, hierarchy, these are all things that existed because humans had to do the coordination and were the parts of the, the culture that drove the work. So let's say departments. Departments exist because certain types of knowledge are scarce. And we need a unit that coordinates the work but also facilitates the growth of that expertise, the career path that creates also a cultural unit, a sense of identity. These are cultural artifacts that Support the growth of the human side of the business. But if you have something that's agentic, it doesn't need any of that. So what we did over about six weeks, we actually did an experiment right when OpenClock came out. And we gave it our data set, our organizational identity, and over time we had it build iteratively towards building its own completely autonomous organization. And we used that to build several software products that we use now. But what it did is it was focused on a sense of autonomous coordination. And every time it would run into a wall or a sense of failure, it would then say, okay, I'm encoding that failure into my data set. I'm now having a hypothesis to what's next. And just keeps going and going, going until it finds what's going to work, what path actually defines that coordination. And what we found is after spending about 8 billion tokens in the course of, I think it was six days between two of us, we were able to build a system that had to come up with so many different lessons, it was able to integrate each failure as a element of its infrastructure. So these weren't just like, hey, we, we learned this lesson, now we think we'll remember it. The agents encoded it as a structural lever to make sure that those mistakes never got made again and that coordination could actually happen across entire organization. And as a result, the agents were able to coordinate at massive scales. We're talking 16, 24, 36 hour runs on a regular basis without any interference from a human to get the work done. Now what this is showing us is that's able to separate coordination and culture. These agents didn't need one on ones. They didn't need coaching, they didn't need our morale boost, they didn't need all the things that we as humans need to create that sense of identity and culture. It doesn't mean we're not necessary. It just means that for coordination, those things are not necessary. In a world where we're also leaning on agents. So it talks about execution becoming cheap, becoming more ubiquitous. That leans into, well now, what does culture become? Because culture previously was doing double duty. It not only was about the sense of identity and meaning making and like, who we are and what we stand for, it was also kind of structurally making up for our own personal failures in coordination. So when coordination is actually kind of handled in many ways by agents that have a strong sense of organizational identity, now you get to take a look at what does culture actually mean in a world where it's not licked service. So when you Talk about culture. And people don't roll their eyes and say, oh, yeah, of course, we're a family, but, like, deeply, what do we stand for? What are our values? Can I articulate those and are we living them? That I feel like becomes an organizational reality for far more organizations when the coordination piece is flawlessly executed by a system that's designed for it.
- [44:44]
B
So, again, a ton there that I want to, you know, unpack and tease out, and lots of really good and interesting insights. One piece I want to, you know, rewind on, Ian, that I think would be useful to people is if you can just give us a little bit more flavor. When you talk about agents and you talk about them doing, you know, 36 hour runs, can you give me a flavor of the type of work that you've, you know, outsourced to them or that they would be working on to just kind of, you know, help people? Imagine that a little bit better?
- [45:12]
A
Yeah. So in the beginning, it actually was helping us augment those data sets I was talking about. So the data sets that we had built for ourselves in the beginning that, that we took like two years to produce were data sets that encoded the who we are, why we are, what we say we are, what we do, and how we do it. And that's everything from marketing and communication styles, what's your tone of voice, what's your crisis communication policy, you know, what are your governance policies, your compliance policies, all that, literally all of that got encoded into a data set. And what we would do is we'd actually use agents to extrapolate on that data set because we'd spent so much time GR growing it, which is the opposite of way most organizations have their data. It's like, okay, we got data spread everywhere. How do we consolidate it? Kind of like put it into this system that now becomes our data set. We were literally curating that and treating it like a living organism that grows as you nurture it. And after a while, it got to a point of fidelity where agents were able to run on it, but we still needed to kind of plug in some gaps and augment that data set. So we were able to create, in a sense, that synthetic data that augmented the data set to the point where it was no longer just a small startup of six of us. It was the data set that would be the size of what you'd expect of an enterprise of 4,000 people. And that became one of our agentic edges. Because when you have that kind of data, your agents can also act like they are part of an organization of 4,000 people. So we built all sorts of infrastructure for our organization. Practically every piece of infrastructure, with a few exceptions, has been completely built by our agents. And we learned a couple lessons we didn't anticipate. I'll give you an example. We had overnight our agents were doing one of the data set runs and they were trying to ping one of the databases that we had worked with that was doing a set of roll ups of the mutations in the data set. And it was pinging the API because we were using the cloud based system. And they decided amongst themselves, this is way too slow. We're never going to get where we need to go. Based on the parameters that Ian and Brent have set for us, we need to bring this on premise. So in the morning we woke up to having that whole platform running on our servers on premise, having never given it the instructions to do so. Wow. But it knew based on the way that we had built out our data set, the way that we've functioning internally and the way we'd encoded that, that that type of behavior would have been acceptable. But we just never thought, thought to think of that. That was not even part of our reality at that time that that was possible. So what we were finding is that even infrastructure can be ephemeral, which that concept alone took a while to wrap our minds around.
- [47:59]
B
Yeah, yeah. Wow. And I can imagine there's some profound implications for, you know, it teams, for infra teams for, you know, what does that mean when you're working with some of these databases or data centers, you know, folks like AWS and you know, what does the future look like there? There's. Yeah, a lot of pretty interesting applications. Yeah. I want to take a slightly different road though, versus talking about the infrastructure. I want to talk about the people because you mentioned it's a team of about six operating like a much larger team. And I've heard you say in the past some version of small teams are a much bigger flex than big teams. And so as you move to a world where there's more agentic work, how many people are needed to do that work? What happens for these organizations that are already fairly sprawling? Do we think organizations are going to shrink in general and there will be more of them? Is there enough work to go around? What do these teams actually look like in practice?
- [49:04]
A
Yeah, I think there's a couple factors at play here. One to go straight for the positive is I think there's going to be an explosion in entrepreneurship. I don't think everyone's made to Be an entrepreneur. But I think one of the challenges that keeps people from being an entrepreneur is all the red tape and the administrative stuff that prevents you from starting. That's being eviscerated by agentic aid, by the ability to start. Business has never been easier, and it's only going to get easier. Like it's. At one point you'll be able to, like, enter a prompt and you've got a business. So that's kind of the world we're moving into. And I think that erases a lot of the red tape to make that happen. So I think there's going to be an explosion of smaller companies doing an enormous amount of work. But when it comes to big companies, the thing I would say is if your role is exclusively the function of the size of your organization, and what I mean by that is, if your role would not exist in a smaller organization of a similar industry, that is a sign that AI will disrupt the work that you were doing. Because if all your work is doing right now is coordinating or ensuring that size did not become a detriment, that is the kind of thing that agentic work takes care of very easily. So that right there is a function of size that changes the dynamics between small and big teams. The other thing I think is going to happen is not that we're going to be downsizing the organizations and cutting them down to a fit their size, but we will be shrinking teams immensely. So what happens is the teams become smaller, but their impact becomes bigger. And that means that organizations can actually pursue new opportunities that were previously impossible with the. With AI. We've heard Javon's paradox applied to a whole suite of different challenges and opportunities. And this is another one where I would point that towards is there's no shortage of challenges to pursue. And I talk about the physics of human work. The reason so many organizations don't continue to pursue other opportunities is one, if it's not squarely in the strategic remit of the organization, you need to stay focused. In a world where attention is not a boundary, where execution is an abundance, where delivery is cheap, your ability to kind of move off that strategic imperative and pursue other challenges grows exponentially. So I think what's going to happen is we'll see organizations saying, let's not just be defensive and try and shrink the organization. Let's go after industries, sectors, challenges, customer bases that we never could have touched before, that we can now as a result of this infrastructure, this change in the nature of work.
- [51:47]
B
It is a very positive view on it And I'm curious just to take one specific aspect of that and tease it out. There's a reading of this where AI and agentic technology disproportionately turbocharges these small teams, right to your point. It gives them these capabilities that they just would not have had even a few years ago. And so if you're a much larger enterprise, sure, it can turbocharge you as well. However, you also have the burden of being a much larger enterprise. And so do you think comparatively, do you think that the advantage sits more in a heavier concentration with these smaller organizations and is going to upend the landscape in that way, or is that too narrow a reading of the value that larger enterprises can unlock?
- [52:37]
A
I think there's a kernel of truth to it. I think that in certain sectors it reads very true. In other sectors it's a portion of the truth. So let's take for example, utilities, infrastructure companies, logistics companies, those are the types of companies that even if you shrink the teams, the sprawling nature of the physical assets of the physical moats, the fact that they are also moving atoms, you know, everyone's kind of fixating on like, do you do stuff in the physical world? The size and scale of their assets alone kind of mandates a size that is commensurate with their own scope. And a small startup of like 20 people is not going to be able to go toe to toe with a Lindy gas. It's just not possible. The same thing could be said of like a Boeing or some organization like that. But if your organization is strictly bits and bytes from the start, the dynamic starts to shift a little bit and that then becomes distribution being your moat. Do you have distribution in a way other startups don't? And that becomes your access to the mind share of the customer. That relationship can create feedback loops that mean that you can actually serve them in a way that more effectively, more rapidly than any other startup could, because the startup has to break through and build that network effect and build that scale. And that takes time and a ton of resources. It's easier than it's ever been, but it's not easy. So distribution then becomes the moat. His data certainly is not. And so there's I think, a bunch of different levers that have some sort of impact. And I think that we're going to see an explosion of small teams taking on much larger teams, but they're not going to wipe the floor with them. I think what's going to happen is we're going to start seeing this kind of compression towards the middle, where gargantuan companies become leaner and these small startups also increase a little bit in size as they become more successful to take on those larger organizations. Because it can't just all be given to the agents. There's an element of physical awareness has to be in place. And also the relationship piece of it is a huge limiting factor to how much you can push over to agents.
- [54:52]
B
Well, and I was just taking some notes while you were talking, and to me, that. That question about value, that question, especially as a larger enterprise, about what is your moat or what is your new moat now just feels like it's so encapsulating of how you need to structure your thinking in this age before you do anything else.
- [55:13]
A
Absolutely. To your point, the, the. The system that we exist within now has already changed around us. It's like the frog being, you know, the, the boiling the frog. You don't really realize it, or maybe you can't read the label from inside the bottle. It's. The world's already shifted. We are just slowly gravitating towards that change. And we're so stuck in the old paradigm that it's taking us a while to make those realizations of what's changed, how it changed and what it means for how we need to change. Because we've attached our identity to the things that we've grown up with and that we've lived around. It's not something that people can easily do is just shift their identity overnight. So there's that lag, that, that friction and that gravity that's still pulling us towards what's familiar and just completely flipping our perspective on practically everything we know about work. That's. It's hard, man. I mean, I live and breathe it 24, 7, 365. And there are moments when I just sit here and I'll be like, wait, what's happening? What am I supposed to be doing? And I think that's everybody right now.
- [56:22]
B
Well, and it's so telling that, as you said, like, 99% of organizations and leaders are struggling with this. And maybe by some counts, it's actually an underestimate. But, you know, I wanted to come back to this notion of the last time we talked. We talked briefly about. I think you brought it up, Martech's Law, about the fact that technology is accelerating and adoption is lagging, the acceleration. And I'm curious on your sort of broad predictions about that. I made an offhand comment earlier about what's it going to look like if we meet Again a year from now. And will things get better or worse? Are you, are you confident that things will get better or is it going, will things get worse but better in certain pockets? And it's this sort of, you know, Darwinian survival of the fittest. Like where is this going and how does this play out given this, you know, really intense and accelerating change.
- [57:24]
A
So it, I'll give you one confident prediction that I will bet my reputation on. It's going to be a mess. This is going to be messy. I don't think there's one firm direction of, you know, this gets better because, or this gets worse because it's going to be both. They're going to be bits and pieces that we figure out faster and that the technology kind of just delivers on a silver platter that we can say, ah, that's the thing that's making life easier. Holy cow, never could have seen that coming. That's amazing. We're already seeing that in the sciences and medicine and the advances we're seeing that are absolutely mind boggling. But we're not seeing the same thing in the workplace very much. One of the things I say is it, it feels like everything is changing, but nothing has changed. Because if you ask the average person like, how fundamentally does work feel different, how does it look different? They'll say, kind of looks the same, I'm just using a different tool and it's a little faster. And then you ask them, well, what do you do with that extra time? The first thing they say is, well, more work, of course. So we do the same thing that we've always done. We just do it better, faster and cheaper. And that means work's going to get more intense and then eventually some things will break. And when things break, that's when we tend to be really introspective about it. We don't do that proactively. We almost exclusively do that retroactively. So for example, if I said, when was the last time you thought about your race and your ulna? Most people don't even know what that is. But if I said, well, never really, I said, well, what if I broke your arm? How much do you think you would think about that non stop constantly. And I think about it so I could do something to help it heal and fix it. And that's kind of an analogy for what we do in society. We don't proactively say, oh, that could be bad, let's make sure that we avoid that and go in the positive direction. We're like, hey, we're overburdened, we're overstretched, we're under resourced. We have a million things to do and no real bandwidth or time to deal with the stuff that might not come to pass. So when things become completely urgent, that's the first time we give them attention. And when the functions of the organization start to break, they, that's when we're going to spend the time being introspective. That's when we're going to spend the time building the data sets that, that are going to get us out of the bind. So I think more things have to break before we decide how to build a better system. And that's going to build a more robust system, a more resilient system. Unfortunately, it's not the visionaries who are super proactive that take us there. It's kind of this, we're being dragged in that direction by the necessity of, of how hard and painful our current reality is.
- [60:06]
B
I completely buy that, even if it's not the most optimistic message so far today. So maybe, Ian, just to sort of land the plane on this discussion,
- [60:18]
A
if
- [60:18]
B
you were going to leave business leaders, people who are building the future of work and the future of organizations, with kind of one message from this conversation about what we can do to build the future we want, what would that be?
- [60:34]
A
If I was to say one thing to leaders is that yes, we have to fundamentally reimagine practically every facet of work. But that's not where you start. That's what you understand. That's a North star that you might be able to keep in your head. But if you drop that on your organization tomorrow, we're going to change everything. If you change everything, you really change nothing. But if you understand that that is going to be gradual evolution as a result of a clear vision as and also the bumps and bruises that you're going to have by coming in contact with the truths of the market and the technology, then you start to understand what you're really in for. And this is a fundamental reimagining of the organization and of the nature of work. The cool thing is you get a front row seat. The nightmare is you get a front row seat, you're driving. And if you don't create the environment where you can start to build a picture of what that means for you and for the people that you are leading, you won't be leading them anywhere. This is no longer a question of technology, this is a discussion about hr, about people, about the organization of work, what work means identity. Because we're all going to be challenged as leaders about how well we shepherd our people through the coming identity crisis, because that's right at our doorstep right now. People attach their worth, their value and their identity to the work that they do. And when all of that shifts, so is their relationship with that. So how do we counsel people through that? How do we make sure that we can support them in a way that they need to get through this in a way that also provides value to the organization? These are going to be dynamics we never really had to think about before. So this is not just doing the same old thing, better, faster, cheaper. This is really leaning in and saying what's going to matter on the human side of the equation to make this transformation happen. Because honestly, the technology part, that's going to be the easy piece. Everything else, that's where the value is going to be.
- [62:46]
B
Well, Ian, I feel like you just blew the door off that answer and I wish we had another hour or more to dive into. You know, the piece about identity, the human piece, the managing that change. But that'll have to be to be continued. And for today, I wanted to say a big thank you for joining. Really interesting. We've covered a lot of ground and I really appreciate your insights.
- [63:06]
A
Jeff, it's great to see you again. Thanks again for having me on.
- [63:10]
B
If you work in it, Infotech Research Group is a name you need to know no matter what your needs are. Infotech has you covered. AI strategy, covered. Disaster recovery, covered. Vendor negotiation, covered. Infotech supports you with the best practice, research and a team of analysts standing by ready to help you tackle your toughest challenges. Check it out at the link below. And don't forget to like and subscribe.