Host Intro (Gene Kim)

I am so delighted about the next speakers, who are Ryan Martens and Melissa Reeve. So just to set some context, I met Ryan when he was still CTO and the founder of Rally Software. He was part of the team who took them public, and I got to attend one of their big room planning events. This must have been around 2015, and it was a revelation. When I say revelation, it was surprising. I learned a lot, but also I was pissed off. I felt anguish because back when, in a certain firm, I was part of these three-day quarterly business reviews, where all the leaders kept on droning on and on. Decisions were already made. It was just a bunch of theater, and it wasn't exciting theater. It was dull, depressing theater.

So at Rally Software, the CEO gave a readout of the state of the business, the goals. The next-level leaders talked about what they needed to do to support it, the help they needed. Mid-level leaders critiqued them, gave them feedback, gave them cautions. And what was really remarkable out of it is that they did it in front of some of their most important customers, which I'd never seen before. And it was also the year that CA had just acquired Rally Software, so they were doing this all in front of their new corporate overlords. And it was just an incredible airing of honest discussion of problems right in front of their acquirers. It was really, like I said, a revelation.

So I've admired Ryan's work for many years, and he and I got to work together with the forum team earlier this week. And they really convinced me that AI creates challenges for the CEO like we haven't seen in generations. And I believe so much the critical capability is: how do you enable distributed experimentation at scale, which is impossible to do if you don't have the right organizational wiring.

So he'll be co-presenting with Melissa Reeve, who wrote an amazing book on this topic about how management roles have to change when the cost of software creation goes to zero. Here's Ryan and Melissa.

Ryan Martens

Thank you, Gene. That was cool. Last two — or three — days have been awesome around here. So we're going to start with an exercise.

I need everybody to stand up. Thank you.

So I'd like you to stay standing if, after this morning's round of talks and discussions, you think this AI adoption thing is going to be impactful to a level of 5X in your organization. If not, have a seat.

If you think it's going to be impactful at a level of 10X, stay standing. If not, have a seat.

Okay. 50X.

We got eight still standing, and most of you stayed at 10X. 10X the change in what happens — whether you account for that through revenue, you account for that through margin, you account for that through happiness, you account for that through human flourishing. This is a consequential moment.

We're here to talk about what do you do with the surpluses from the real gains in this adoption cycle. Our question is, though: why are so many organizations treating this like a tools rollout?

We have an opportunity for 10X to more, heading towards 50X over maybe not an 18-month time period, but we'll show you a couple of companies who over the past 10 years have taken something to an amazing extent.

And that's really where Melissa's book comes into play. Melissa has compiled a set of case studies over the last few years of organizations rolling out this adoption from different places in the world that'll show you this thing is the most tsunami-like event we've ever seen coming before.

So one of those case studies started in 2008, and it was one decision made by a company.

And 17 years later, that organization had gained 240 million customers. We're going to come back to that company.

But first, we're going to talk about your options for that surplus. Vibe coding is producing real gains, and the question becomes: what do we do with those?

We think you've got three choices.

You can harvest those gains. There's plenty of financial gravity pushing down on all organizations, especially in the Western Hemisphere. It's very tempting to take those gains and put them straight to the bottom line. It's also very tempting to not really even manage or understand those gains and let them slide down there by themselves.

Our suggestion would be that you move past experimenting and over to building — and that is being able to reinvest those gains into some of the stories that we've seen today around common infrastructure, around how your employees come along for the ride and skip the spot in the middle, where I think Melissa would tell you there's a lot of people who've kind of gotten lost in this transition.

Melissa Reeve

Yeah, and if you don't invest in the building, you might decide to invest in experimenting, your R&D. And that's great, but it often results in motion without progress. There's no milestones. Ryan and I have joked it's like dinking around with AI.

So we're going to push you towards building, and we're going to push you up in this stack. One of the things — if you can't make the jump to full-out building, and you can't have the conversation with the CFO yet, and you're kind of protected in your environment right now — at least you can start with the CFO in an "I trade you" situation.

So back in 2005, we became B Corp certified. That allowed us to get an 80% discount on our Salesforce licenses. Thank you, Marc Benioff. I went to the CFO and said, "Jim, I'll trade you. I'll give you half those savings — you can have them — but I'm taking the other half towards corporate social responsibility and our citizen engineering efforts."

That is a simple little discussion that had — he pushed, because all CFOs push, but he very quickly decided that was a great deal. That's a way to take some of the savings very simply back to yourself and be able to keep investing in what's going on.

Ryan Martens

So we talked about this 18-month window, and what can happen in those 18 months. So the last time we had a transformation like this — or an impactful technology — was with the internet. And many companies, when they were getting started, simply took their offline brochures and catalogs and put them online to create an online brochure or an online catalog.

But there was an exception to this, and one of those exceptions was Amazon. And for 26 quarters straight, Amazon decided to reinvest their profits to reimagine the entire customer experience. And of course, we know where that ended up.

We have the similar opportunity now with AI — to not just recreate existing processes, to reimagine what is possible.

Do you know where your last efficiency gains went? Anybody? We've got one hand.

We spent two days in the forum group talking about developer productivity, trying to measure outcomes versus outputs — Mick's new book on outcomes versus outputs. It's not easy. Totally understand that. But it's your job to figure out how to do this.

If you're going to have the conversation with the CFO about this stuff, you've got to get your arms around what's going on here so that it doesn't kind of fall forward like Parkinson's Law — and that is, we'll just expand the air inside the project or the two-week time box or whatever, and we'll fill it up with a bunch of other stuff. That's the easy thing to do. It's the thing that'll just happen, as opposed to cutting the time box back to four days, like we've heard in a number of discussions in here.

When you cut the time box back to four days, now it becomes a lot easier to figure out where those efficiencies are coming from, in the form of cycle time and less waiting time and less waste in the process.

That's what's in Melissa's book. Of course, it's an IT Revolution book. It's coming out this month. It's the story about how to get your way there.

Melissa Reeve

So one of the ways you can reinvest is in the dynamic routing that Gene was alluding to earlier today. And many of you have started to build this routing. The hyper-adaptive model is within the book, Hyper Adaptive, and it starts to suggest a blueprint for moving your organization from where it is now — what I call a linear organization, where we have strategy to execution, we have concept to delivery — into a more AI-native space, one that can sense and respond in real time.

What we know from our previous journeys is that this doesn't happen overnight. We also know that it takes deliberate effort and deliberate infrastructure. When we put PCs on everybody's desks in the 1990s, we didn't just leave them and say, "Go play with them." We spun up IT help desks. We spun up all kinds of support. And that's what the hyper-adaptive model starts to do. It starts to put forth all of these support structures like AI leads, AI activation hubs, AI impact hubs to help the organization rewire itself.

Again, we've heard many stories of people who've started down this path over the past two days. This is the system that pulls it all together.

So we take a look at nine different dimensions across these stages. We've got ten up here because I've given you a bonus. Because what we've heard over and over again is that the bottleneck shifts. And these are all the places that we also need to address as the bottleneck and the entire system starts to rewire itself.

How does AI impact budgeting? How does it impact communications? How does it impact the roles? All of these are things that we need to be taking a look at.

And this'll be fun because this is a build.

So one of the things we've also heard is the importance of learning and feedback. And so these structures — this dynamic routing — starts to create what I call a bi-directional AI learning flywheel.

So let's say that you have a new insight that happens on the front line. That insight is captured by your AI lead, who then hands it off to an AI activation hub to codify and document it.

So that activation hub then shares it with the network of these activation hubs and localizes it for that context. Imagine you have an activation hub in legal or in finance, and the activation hub also encourages that pairing that we've been talking about, where you have somebody who's really expert at building out an AI agent or automation combined with that local subject matter expertise.

So we scale it, and then we sustain it by embedding those learnings into an AI knowledge engine. And in this way, we create that routing.

Let's go the other way with the AI learning flywheel. Now imagine there's a new AI capability. We route that through our AI activation hubs. And then the activation hub puts that new capability into the hands of your AI leads, who then get it to your frontline practitioners. And in this way, we're creating feedback loops on the learnings, and we're distributing the new capabilities throughout the organization. This is that dynamic routing for your organization.

Ryan Martens

So we alluded to an organization that implemented this type of system in their organization. They started their journey in 2008. That organization was Ping An Insurance.

What's key about Ping An Insurance is they started with their data infrastructure, and they got their data house in order. The next thing they did is they put their customer at the center. This started simply as an insurance underwriter, but because they really wanted to service the customer, they expanded it to healthcare. They expanded it to finance, and they created an entire ecosystem where, if they knew you lost your job, they might offer up mental health opportunities. They knew that if you were a good driver, it might also indicate that you were a healthy individual, and they could adjust premiums accordingly.

And this had measurable results. Those who operated within the ecosystem had over a $5,000 share of wallet or contract, versus those outside the ecosystem — other insurance underwriters — had a fraction of that.

So our proposition is that what was, in 2008, Ping An's decision is now your decision in 2025. I would propose that this is actually a business decision, not a technical decision. Your job is to push this up to the business level. If you're going to rewire the organization, it's the only place in the organization where that can take place.

Many organizations get this backwards, and that is that they're trying to bring the capacity and capabilities to the organization and show them, "Here's what it can do," when what we saw from Ping An was the business coming forward and saying, "This is the customer we're looking at. This is the value we're trying to bring to that customer set." A whole other level of experience and capability than single individual entities — in insurance and healthcare providers — could give, but something that was complete for their entire life. Start there.

What is the purpose of business? The historic purpose of business has been human flourishing. We've gotten confused in the last 20 or 30 years with a whole bunch of people who did not enact any legislation. There are no laws, there is nothing in the books that says shareholder primacy is the way this has to work. That is financial gravity pushing down on top of a late-stage capitalism world.

Your job is to be able to bring this back to: who do we serve? Do what you love to serve the people who love what you do.

That's the question here.

AI changes what you can do, and it changes who can do it. That's the really interesting place. As the roles shift, as the way we start to work together — how we work together changes — we open up what none of us know what is going to really look like. Ping An has a vision of where that's at and is operating that effectively today. But for all of us, we're going to end up in very unique places.

I think the recent Claude Code Hackathon — this is a global hackathon — is a case in point. The top five winners included a lawyer, a cardiologist, a musician, a civil engineer, and one software developer.

What does that tell you about the creativity that's about to be unlocked in your organizations? What does that tell you about human flourishing?

We're at a watershed moment.

We're going to switch gears for a second. We've been talking at an organizational level and ecosystem level — however you want to describe it, the top level, the third level in Gene's wiring, the rewiring the organization level. But there's something just as important happening inside you, especially as the leaders of this transformation. You're the model of how this has to run.

And I'm going to tell you, you need to be the best you've ever been. And it's an opportunity for AI to amplify you — but let's have it amplify the best of you.

So I'm going to tell you a little bit about my story last year. As Gene said, I retired in 2015, and I started doing this. I raise cows in central Colorado on a regenerative ranch. What we focus on is healthy soil. Healthy soil drives plant diversity, which drives better feed quality, which drives healthier animals with no vaccine and no inputs, which drives proven better meat for humans and human flourishing.

The same message goes for what we're working on here for you as an individual — and that is focusing in on the core of what's going on that's going to create that impact is the critical component.

I started my journey in coming back into this world with a partner who pushed me hard. I'm fortunate enough to be working with a gentleman who helped create the d.school at Stanford, was on the founding faculty, and has spent the last 10 years going around the world helping bring design thinking to organizations all around the world.

So when he started pushing on me — a civil engineer, a computer science engineer, as well as an MBA — to round out and be the best I could, I started with marimba. My second step was African singing. My third step is I picked up Julia Cameron's book, The Artist's Way, and started writing morning pages. Because if I'm going to channel what's coming to the universe through me, I've got to have a practice that can do that. I have to have a way to balance what's coming in and not overload myself. I have to be able to manage my day with slack. And how you move through your day determines whether you build slop or super high-quality stuff.

This is the same message that you've heard all morning about the organization. You need to take that same discipline and practice to yourself.

I started building with John. John's built a whole bunch of — we've got some watch apps that put you in control of your agentic teams from your watch or from your phone, that store your data in a sovereign data store, and that crystallize that work so that you can pass stuff on to your agents in a way that separates tasks from design intents.

I started building the piece of our product that would then stage that work into third-party AI tools and building tools. And believe it or not, I started from a picture of, "Hey, this is kind of like Rally, but it's for individuals." So I started building something like that.

And then I did the crazy thing that most product managers really don't like to do — I went out and talked to a bunch of people and said, "What do you think?" I got humble pied, as most people do, and I got some feedback, and they said, "It's too much, too fast, too hard to use, too much work." So I radically simplified the thing and built a guest mode and a much simpler way to get into it, and get to components and prompts that you could feed to the AI more effectively.

Then I went to a place where my LangGraph agents got way overloaded and expanded, and they couldn't reliably write JSON to the database, and things were going haywire, and it was not going in a good way. We were kind of spiraling our way down. So thank God Claude Skills came along, and I built a whole bunch of skills to do exactly the same thing. And then I decided, well, I need a UI to control those skills, of course. So I built the UI that drives — now pulling in specs, creating changes to those, analyzing them for holes, for risks, for where's the soul of your product — and built that to start changing things.

And then I went ahead and did something crazy and sent my design spec — you're always looking for something to test — into my skills. And what did my skills tell me? They said, "You have a giant risk. Many of your users who are in CLI environments are never going to open a UI."

So this was kind of an in-your-face mirror moment. What happens with AI in a lot of cases is it shines a light on your radical assumptions. So I'm open-sourcing all those skills, and I'll give you a QR code to download all those. They've been through now six months' worth of iterative development through my LangGraph agents. But it's basically a product management set of skills that you can add to compound engineering, or you can add to superpowers as an augmented piece for new development.

That's the second change that I had to do as a person.

The third change is you have to realize we're all on this journey. We've heard about people who are starting in a skeptical place — whether they're mid-level engineers or whatever role — and they have to learn for themselves how to get to a place where they can start experimenting safely and feeling like they're not embarrassing themselves, basically.

Most of us are focused on this middle cycle of being productive. I'm going to tell you, if you're going to get amplified by AI, that's a false finish. You've got to figure out how to pull the best creative capabilities out of you to drive the future intent that you need to solve the problems for your customers.

The last place we're going to get to is a place where our flow cycles look like design, they look like development, they look like learning, and they look like updating the way you work — simultaneously, into our cycles. That's what generative users look like using generative AI. That's where we're trying to get to.

We have to be empathetic that all of us have to go through at least three or four major changes here. Shift left, for sure. Becoming the chief of your own agentic team. Move up to a higher level of management and orchestration. But you also have to go through this fundamental learning journey where you can't shift your way along that path. You need space, slack, and time to do this without blowing up.

We need you to not blow up. You have to lead this journey.

One thing's for sure: AI is never going to be better than AI with humans who are being amplified by it. I don't think we have to worry about all the proselytization about it's going to take over the world. We are always going to be in a better place as long as we can channel what we're doing through us effectively.

If you want to find what we're doing, you can follow us on Substack. You can download all my skills from my repo. You can contact me directly. I would love to have five people who wanted to push the notion of how you amplify yourself forward to contact me and jump in on some of our earlier release tools.

The question becomes: with this surplus, what do you do with it?

Do you simply reduce headcount and harvest the gains?

Do you invest in your organization and the dynamic routing that's needed to create learning flywheels?

Do you invest in yourself and your own human flourishing?

We, of course, would encourage you to do all three. Not only harvest some of the gains, but invest in your organization and invest in yourself.

So we leave you with three actions that you can take. One is share the blueprint — share how to create this system that responds to AI and moves your organization to an AI-native place. Invest in that wiring that we talked about. And amplify yourself.

If you do that, you can sense it — I know you've felt it. The aha moment that you got along the journey: that was you amplified by AI. Your goal is to create that multiple times a week.

Let's execute towards human flourishing and towards AI that amplifies it. Thanks.