Chris Blackburn

My name is Chris Blackburn, founder and CEO of Liatrio, and I'm joined on stage with VP of Innovation of Liatrio, Robert Kelly. I'm really excited to talk to you all about how we are seeing and expanding and scaling AI throughout the enterprise, so that it's not just for engineering talent and the delivery or the development function, but across the entire value stream.

Real quick introduction to Liatrio: we're a boutique consulting firm. We focus on helping large complex organizations navigate their technology transformation. I was a huge proponent of the DevOps movement and really excited about how everything moved along through this. What's unique about us is we are enablers. We help the organizations become a better version of themselves. That's shoulder to shoulder, fingers on the keyboard, guiding them along this journey and kind of going into battle together. We work with some of the greatest brands on the planet. It's exciting to see them on their transformation journey overall.

We look at ourselves as guides, not the heroes. The customers should be the hero and ultimately own their transformation for the long term. We want to just be part of igniting that future for them and making them the best place on the planet to work, to grow and retain their awesome talent.

One other thing I want to talk about is this community. This conference has been so huge to me. Funny enough, I actually met Robert at the very first DevOps Enterprise Summit in South San Francisco back in 2014. Somehow I convinced him to join a very small but growing Liatrio about a year and a half later.

It's so important to me. I've built such an amazing community with everybody here and learned so much from so many folks that are now coworkers, customers, and a lot of friends that are here. Any recommendation to you all is to really take this journey of this community and use it for your benefit. We're coming into what I think is probably the most pivotal part of all of our careers right now with the AI movement. If I want to go do something difficult, I want to do it with others. I want to do it with community. I want people to challenge me and ask questions, give me advice, and give me input. There really is no better place to do it than this place right here.

Yes, the talks are amazing, but please make sure that you build the relationships that you're going to take away from this conference for discussions to happen in the future to come.

All right, so let's jump into it, starting with an analogy: the modern tractor. It really is quite the technology evolution, for certain. The new tractors today are absolutely amazing. GPS-controlled, can do so much more than they could before. You could ask our friends that are here at the conference from John Deere; they'll tell you way more about it than I could ever pretend to even know. But it sure is exciting to see the journey that the tractor has been on.

I was watching a YouTube video about the latest and greatest tractor that's completely GPS-controlled, auto-steer, and all that type of stuff. They start calling the driver of it "steering wheel holders." It's all they do: just hold the steering wheel and make sure you're there in case something goes wrong, because the machines are going to start handling a lot of it on their own.

This is awesome, but this is farming technology, right? Agriculture has taken 150 years to get to where it's at today. What do we think we look like when we get there?

Raise your hand: who's heard of the 10x engineer? I'm sure most of you have. I think Robert would kill me if I didn't say that we are hiring, so if you know 10x engineers, please send them Robert's way. He would love to chat with them. But have you heard about the 10x farmer? Anybody? 10x farmer? No, maybe not.

I think the reason why you haven't heard about the 10x farmer is because farmers are 10 times more productive now than they have been in the past, but it took a long time to get there. It was a long journey. Back in 1900, for example, compared to today, we produced five times more agricultural output than we did in 1900 with almost 90% less labor. The technology has taken us on this journey along that way, and we've changed so many things about the entire process around farming to get that food from the farm to the ultimate customer along the way.

It isn't just about taking that tractor and giving it to a farmer. You need to improve the entire ecosystem, the end-to-end flow of that value from planting the seeds and nurturing them and growing them all the way to picking them, putting them on tractors and trailers and all that type of stuff, and getting them out ultimately to the customer.

Mechanization was a big move here: bigger tractors, more farm equipment to help plant seeds, and all that type of stuff. Transportation services: in 1900, there weren't highways, a lot fewer railroads, no big rigs, no things like that that allowed you to move the stuff that you cultivated. Storage, either cold storage, freezer storage, all that type of stuff to make sure we can keep that agricultural output. Irrigation alone: there's been a 20x improvement in how much irrigation we use on the crops to grow them. Then obviously processing and distribution networks: the mills, the conveyor belts, and all that type of stuff.

The point here is that it's not about just giving the farmer better technology. If we just gave a farmer that tractor from today in the 1900s, they wouldn't produce anything more than they do today. They would produce the same amount because the rest of the ecosystem is not there to support it. That's what we're going to tie into our talk about the rest of the ecosystem.

Ultimately, a final note on this is: back in 1900, the average farmer provided enough food for approximately 13 people per year. Today that number is well over 150 people. There's your 10x farmer compared to 1900. However, with such a slow evolution, it wasn't a transformation. The AI journey is going to accelerate that time to turn us into 10x engineers and ultimately produce that 10x outcome and value because of the entire ecosystem end to end. Robert, take it away.

Robert Kelly

Awesome. Thanks.

Okay. So what does AI transformation mean for your organization? Not much. The entire value stream and enterprise has to change, or we can't take advantage of these new capabilities. Until now, we've relied on individual contributors to adopt new technologies. Really, the enterprise slowly adapts or accommodates these things over time. You may have had multiple transformations, multiple DevOps transformations or agile transformations, but there's always a corner or an area of the organization that's really left untouched, or they're not taking part in that modernization journey. There's always somebody left out.

But this time with AI, it's different. It just is different. There isn't an area in enterprise technology or the business that's going to be left out. Really, to make that successful, we have to all see ourselves in that transformation.

Of course, there are going to be completely new tools. There are dozens of new tools. Even though that's how we were introduced to AI today, you may have been introduced to AI through ChatGPT, GitHub Copilot. Really, these are just how we interface with this new technology. But we need to avoid the trap of just thinking of AI equaling tools. That's really one of the bigger takeaways I think we have today: it really is not about the tools.

What really matters here is that everyone's work, our workflows, have to change. We have to be able to ask anyone in our organization, "How is AI impacting your work?" And we should be able to see that impact. It should be more than just, "This is how it's impacting me." They should be able to show us how it's impacting us.

With some of us, it's impacting us sooner and maybe more drastically. Software engineers, for sure, are being impacted sooner and more broadly today than any other part of the organization just yet. But there are going to be others that really, I'll say, have to be fast followers on this journey immediately: all the folks involved in aggregating data, building data infrastructure, modernizing data platforms. This is going to be everybody's responsibility. If you hadn't thought that data is your responsibility, guess what? Data is the fuel for AI. It's going to be part of this transformation.

I'm sure we have some governance and security folks here today as well. One of the bigger changes here is we've got to be active in this change. We've got to see ourselves in this transformation, making ourselves available to the organization as accelerators. Ultimately, that's going to mean culture change. This isn't something we're going to snap our fingers and say, "The culture's changing." It's really just got to be something that's very intentional. We're just going to see ourselves in this change. It's a huge change that we all have to see ourselves in to make possible.

So what does this actually mean? From an end-to-end AI transformation perspective: new tools, workflows, rethinking roles and team structure, new platforms, and innovation product thinking for the entire organization. That's new for a lot of folks. Really, we've got to stay connected to the changes of the industry. As new AI advancements come out, we've got to be really attached to those and understand what those mean.

Digging in a little bit here on some of these actual changes, individual workflows are being impacted more than anything else today. So absolutely, new tools. The biggest change here is we've just got to be more product-focused. We've got to move from thinking about individual tasks to thinking about more product outcomes and what that means for the organization. That means our individual workflows are going to be changing pretty drastically.

If the past couple of years have been about getting new tools and licenses for developers, we're moving past that. All of these tools are great. Whether our developers are using Claude Code via Copilot, Windsurf, Cursor, these tools are all great. Ultimately, if we're attached to the latest frontier models, they're going to be able to generate most of our code. Almost all of our code can be generated through AI today, but really only if our workflow changes. That's just got to happen.

We've got to move from thinking about AI as a helper. This is a big one. I'll say this, and you can talk to me about it later: we've got to move from AI helping you to you helping AI. That's a change. We've looked at AI as an assistant. It's nice. It's a pair programmer. But really it's our responsibility to guide AI today.

That really means rethinking our workflows at more than an individual level. We've got to bubble that up, really bring that up to the team level, because almost zero, probably zero instances, is an individual responsible for delivering value today. It's always going to be the team. The team's responsible. So if we improve a single developer workflow, how does that translate to value?

Again, it means we're going to have more cross-functional roles on the team. That's one of the bigger changes here. We've got to understand we're going to have more agents. We're going to be able to do more. This extends to everyone in the organization. In order to really see this change, you're going to know it's successful if a team shows up on Monday and builds five prototypes with a product owner in the room and ships something that's better than they would have otherwise by Wednesday of that week. It's even better if the product owner shows up with a working prototype for the team. That's going to be a big change.

Again, I see lots of change in the team structure. I think the teams are going to become much smaller. I see it more than I think it; I know this is what's happening. Teams are becoming much smaller, and that means we're defining more products as well. It is going to be a big change for us to really think about productivity and reimagine productivity as product outcomes for the individual. That's one of the huge changes here: an opportunity to really think about how we're aligning our teams. If you don't have stream-aligned product teams, that's pretty much a must-have right now.

At the enterprise level, that really means we're focusing on innovation and having a product mindset across the organization. Again, everyone's got to be part of that to make sure this is possible. We've got to bring all stakeholders into this change.

Where we have been in the past is a lot of, "We've got our service organizations, we've got our organizations that provide services to the rest of the organization as sort of maybe an afterthought or upstream or downstream." Security and governance is after development, or testing happens after development, or product design happens before development. We've got to bring these folks together. There's a need for active collaboration more than ever.

That means that security, for instance, can't be the org of no anymore. We've got to be active in this. Anybody here in the security organization? Good? No, because really, you're going to come last anyway. It doesn't matter. No, it's fine. The goal here is saying, it's actually not good. Where are the security folks? They really need to be part of this. So next time, invite them, please. That's something we really see we have to do. We have to bring the security folks into conversations early.

Right now, there is a potential that they're looking at these new capabilities as challenges. We've got to help them through that. We've got to work through this together.

Again, I think Gene mentioned this early in the opening remarks: with the AI transformation, we're all going through this together. We all have the same challenges. We're all technologists here, so we have responsibility to stay informed about the change and be drivers in this change. This is a great opportunity to align on all the common success stories, learning to innovate together and really take advantage of our community here. This is super important. ETLS is a great opportunity for learning together and building what we want to see, this next level of consensus. Take the opportunity while we're here to have all these conversations. Especially if you see something I'm talking about, I'd love to talk about it further later.

I'm going to get into a little bit of the detail here. This is where I'd like to maybe challenge some of the folks here. The majority of the enterprises that I speak to and work with, and really just anyone that I've seen, are in this phase. I'd like to say in this box on the left here, that's yesterday. That's really where there's no consistent AI-native workflows across individuals or teams. The discussion there really is about ROI: who has GitHub Copilot, or in many orgs still arguing about which tools to purchase, really to get past that. I guess a lot of organizations are building their own "my company GPT," their version of ChatGPT internally, using subpar internal models. They actually suck quite a bit. They're not that great. We might see a 10% productivity boost, but that's not translating to enterprise value.

We need to make sure that we're focusing on bringing the entire organization up to a good level of AI fluency. That's where I think we should be today. I have this box highlighted here on the screen where everyone needs to be closer to becoming AI-native.

I expect, and we should be expecting, that AI is generating 90% or more of the code that we're delivering. That's a big number. Maybe outside of this talk, I'll go ahead and say it: I expect it to be somewhere close to about 400% or 500% of the code that we're writing, because we're writing so much more, we're able to deliver so much more. It's not just 90%. It's an order of magnitude more than that potentially.

That means that these new workflows are going to require a lot of change upstream and downstream. If we're building 2x or 3x more, that's going to apply some pressure to the rest of the organization. That's where I think some of these big changes have to happen at the individual and the team level, or we won't be able to move to the next phase here, which is making AI asynchronous, bringing these capabilities to the rest of the organization.

If we're stuck in this synchronous workflow where every developer has their own workflow, they have their own process, and you have one developer that's writing 10x more code than the next, that's just not helpful. We've got to make sure that this is visible, that this is actionable, and that everybody understands how this is happening across the organization.

Ultimately, I think that ends up looking like we're building agents, we're building these AI-enabled products, not just for our individual workflow, our team, but the organization. If we have our teams building these products, then it's going to be available to the rest of the organization. But what's more important is what your organization is providing to their customers. If we don't improve our individual workflows, bring it to the team, have a common approach across the organization, we won't be able to provide these new capabilities to our customers. There's going to be a lot of change here. AI agents really are going to be a huge factor in commerce, and we're not going to be able to build those products unless we start from the ground up and enable our organization at a high level.

Let me get into a little bit of what this takes. I mentioned AI should be generating 90% or more of our code. This is where vibe coding has evolved. It's not just, "Send in a prompt and expect to get something we can ship." This is an actual change that needs to happen. This is the workflow that I expect everybody to be following today, because this actually works. You'll see more of this across some of the talks and some of the folks that are really moving deeper into this process.

This is what I like to call that we're entering the requirements documentation renaissance. You all have worked with the requirements docs throughout the years. But that's not waterfall, and it's probably not going to be Scrum either, by the way. Sorry about that. This means everything moves from individual tasks to becoming outcome-focused. There's a huge opportunity to get in here, get deep, and understand what this workflow looks like. This is the spec-driven workflow. If you haven't heard about it, I'd love to talk more about it.

This is where we're democratizing the work across humans and AI agents. If we don't follow this process, we're just not going to be able to level out that playing field of how work gets done. Just to give a little bit more context, this is where we're breaking down a PRD and guiding agents as a human in the loop, or we're going to make sure that the agents have what they need to do the work asynchronously.

If you've heard of context engineering, you'll hear that a thousand times during this conference, I'm sure. That's where all of this work matters. If we're not building a common context or a common language for humans and agents to work, then it's not going to scale. Again, it's important that this happens not at the individual level, but at the team level and at the organization. That's where this needs to happen.

It's got to be done collaboratively, and we've got to use AI to help us build these well-documented product requirements or stories or tasks so that agents can work on them. Right now, you'll see some 10x engineers or cracked AI developers using Claude Code. They've got a lot of agents they might be running on their laptop, but who sees that? Who sees that work? "I go and I cranked out 5,000 lines of code with my 10 agents. It's all well tested. It's secure."

But one developer doing that work is not going to translate to enterprise value. It's just going to gum up the works. We're going to hit so many walls, and honestly, that developer looks like a jerk today, even though they created a lot of great code. Again, our workflows are going to change almost entirely. We really need to stop thinking about this as a tool. I mentioned Claude Code, but any one of these AI-native tools are really pretty great. We've got to get away from that and really focus on the workflow.

Chris Blackburn

Nice job, Robert. Thanks.

I got a few slides to close out here with. I'm sure many of you have seen this one before: the graph of the growth of the administrators to practitioners or physicians inside of the hospital. It's insane, the number of administrators. People that are helping the process more than anything, or trying to help the process more than anything. In a lot of cases, we see them getting in the way.

I believe that this slide right here could say enterprise technology on it, if we want to have engineers on the bottom and project managers, program managers, managers of all sorts, managers of managers, with the higher line there. I think it's really important to bring people closer to the work that is happening. We want to make sure that the people that are writing the code, even if that's humans or agents, that's where the value is actually produced. Everything else besides that is either overhead at best or waste at worst. We need to cut waste from the system and we need to make sure that overhead is accelerating software development and the creation of those new products.

A lot of things that Robert talked about, creating that digital value and things like that, is smaller teams working much closer together. Business and product are not at the beginning of the flow and operations at the end, but everyone functioning there together.

Again, like we talked about, I actually stole this from the chief product officer of LinkedIn. He has a really cool post online. Please go check it out. His name's down there at the bottom, talking about this flow of value here. For some reason, we are all focused on the writing of the code. It's the software developer and making them more efficient.

All of our studies, all of our information that we've found, says that the smallest amount of time and of human capital, and the smallest amount of financial capital or cost, goes into the coding effort. It actually happens outside of that. So we're sitting here right now trying to accelerate the coding function of it, the software development function, and forgetting the rest of it.

While AI is helping and accelerating that today, we need to think about AI across the entire value stream, the end-to-end portion of it. That's what we're talking about. It's going to completely change workflows, it's going to change roles and responsibilities and how people interact together and work together as a team. Again, if we take the farming function in here as an example, just giving the farmer a much bigger product, a better product, a better tractor or something like that, is not going to work. End to end, we've got to think about it all together. How to apply AI to the entire value stream.

Closing this out, it's time to action for sure. Let's get to work here. The vision here and the future of it is AI transformation is truly underway. You're either going to be part of the journey when we know that the future is not fully solved yet, or you're going to wait for others to do it. That's really concerning to me around waiting.

There is no playbook. There is no one thing you could pick up and say, "This is the way that's going to work and this is how we're going to operate." You've got to build the playbook as you go.

The technology is here, it's here today, and it's improving all the time. Take the technology and work alongside it. There are standards, there are platforms, there are things to build upon, but they need to be put in practice inside of your organization and how it operates.

And then the competition. The competition is coming. I think there is going to be a big wave of organizations, big, small, Fortune 500, whatever it is, that are going to struggle underneath the anti-competitive nature of not moving forward in accelerated software development. We've got to create these digital products and digital value faster. Moving everybody to be thinking product mindset and product function is such a huge deal.

Let's go. Let's get together. Let's do this. Again, reminder: this is a community. Please build the community. It's absolutely amazing. We would love to talk about this more. Robert has his innovation team here. We've got a booth, booth number nine. We're going to show off a lot of the demos and really awesome things that we're doing. Please come say hi to us, check us out on LinkedIn, let's carry the conversation forward. Thank you all very much. Really appreciate it. Thanks.