What Happens When You Win with Vibe Coding?
Your Efficiency Win Creates a Decision, Not an Outcome
Last week I was on stage at the IT Revolution AI Summit in Denver with Ryan Martens, Director of Manifest AI, and he opened our session with a simple exercise. He asked the room to stand if they believed AI would be a 10x multiplier for their organization. Most people stood. Stay standing if you think it could be 100x. Some sat. Stay standing if this could be a 1,000x moment. A longer pause. A few more sat.
Then he said: look around. You just told us you believe this is transformational. So why are most of your organizations set up to capture none of it?
The room got quiet (in a good way).
What followed was a conversation that felt different from most of what I hear at tech conferences right now. We skipped the ‘wow’ moments of AI and talked about what happens after the productivity gains arrive. And based on the questions we got and the conversations afterward, that’s a topic people are genuinely hungry for.
So let me share the core of what we covered, because I think it applies to most of you reading this.
The Surplus Is Already Here. Have You Decided What to Do With It?
Vibe coding and its concomitant productivity gains are real. EY’s most recent AI Pulse Survey found that 96% of organizations investing in AI are experiencing some level of productivity improvement, with 57% describing those gains as significant. Developers are building faster, freeing up time (although I think the backlog is so deep, it may or may not feel like real progress).
In some organizations, headcount is being reduced (though the EY data is interesting here: only 17% of organizations seeing productivity gains actually used them for headcount cuts. The headlines rail otherwise… so much hype.
The gains are landing. So, the question becomes…what happens next?
In both the research behind Hyperadaptive and in the conversations I have with enterprise leaders, there are essentially three postures organizations take toward an AI-generated productivity surplus. Ryan and I presented these last week:
The first is Harvesting. Every efficiency gain goes straight to the bottom line. AI becomes a cost-reduction tool. The CFO is briefly delighted. Headcount shrinks or hiring freezes. Fast track to irrelevance, because your competitors are doing something different with their surplus.
The second is Experimenting. A few power users are doing genuinely cool things. There’s energy, there are demos, there are Slack channels full of prompts. But there’s no map, no milestones, and no organizational learning happening. Motion without progress. (I see this one the most, honestly.)
The third is Building. The surplus gets deliberately reinvested in capacity. AI Champions get protected time. The organization starts to rewire itself, one stage at a time. Compounding advantage.
Be honest with yourself about which one describes your organization right now. When did your organization last have an explicit conversation about where AI productivity gains are actually going? Can those gains even be measured?
The Missing Piece Can’t Be Found in AI Tools
The 2025 DORA research found that top-performing organizations are seeing 20 to 60 percent productivity gains from AI. Most organizations sit at 5 to 10 percent. The tools are the same. The models are available to everyone. The gap reflects differences in human factors, not access to technology.
The conundrum faced by most organizations is that coding speed is up dramatically, but the value captured by most organizations has not moved at the same pace. Because capturing the value requires a deliberate decision about reinvestment, and most organizations have not made that decision explicitly.
There’s a useful frame from Eric Ries’ work called the Temptation to Harvest. When a new capability creates genuine value, leaders face a choice to extract that value as margin, or reinvest it to build something bigger. AI has handed every organization a surplus. And right now the default choice — the path of least resistance, as well as pressure from the board — is to harvest.
Ryan made a version of this argument from his own experience, describing an earlier fight with a CFO about reinvesting software licensing savings into corporate social responsibility rather than letting it evaporate into the P&L. The CFO wasn’t thrilled. The reinvestment kept the organization focused on purpose and impact. That’s the same choice in a different context.
Related Article: Why SaaS Companies Require A Bigger Mission
One Decision, Compounded Over Seventeen Years.
The clearest illustration I know for what deliberate reinvestment looks like over time is Ping An Insurance.
In 2008, Ping An made a quiet decision to make data the foundation of everything. Not AI specifically, as they didn’t have the AI yet. Just a commitment to getting their data house in order and organizing around the human being in front of them rather than the products they were trying to sell. They restructured around the customer. They established a dedicated technology division, not as a support function but as a strategic driver. By 2017, they had formalized AI as one of their five core technologies and committed roughly 1% of revenue to AI research and development annually.
Today, Ping An serves over 240 million retail customers. Customers inside their digital ecosystem hold an average of 2.9 contracts each, compared to 1.2 for customers outside it. Revenue per ecosystem user reaches $5,288, compared to $1,399 for non-users. Nearly 4x.
That’s the result of one deliberate decision made seventeen years ago about what kind of organization they wanted to be, compounded. The technology followed the answer.
Ryan put it well on stage: Ping An’s 2008 is your 2025.
Related Article: The Five Stages of Becoming AI-Native
The window for this kind of decision isn’t what it used to be. Unlike the decade-plus runway organizations had with digital transformation, research on AI competitive dynamics suggests the compounding advantage for early reinvesters is closing much faster. (I’ll hedge that slightly: the exact timeline is contested, but the directional pressure is real and consistent across multiple sources.)
Three Things to Carry Out of This
I closed our session with three asks that I’ll repeat here, because I mean them.
Invest the surplus intentionally. Before the next efficiency gain disappears into the P&L, have the explicit conversation about where it goes. In everyday budget conversations, with specifics.
Name one AI Champion. One person in your organization, protected time, actual mandate to spread what they learn. That’s where the 20 to 60 percent gains start, and it costs almost nothing compared to what you’re already spending on AI tooling.
Related Article: Why Appointing AI Leads Isn’t Enough (and What to Do Instead)
Find a way to amplify yourself, not just your teams. Ryan’s part of our session touched on the human journey. He talked about the path from curious to generative to amplifying yourself with AI. I’ll let Ryan tell that story on his Manifest Substack (mnfst.ai), because he tells it better than I can. But the organizational and individual journeys are connected. You can’t build an AI-native enterprise without leaders who have made that journey personally.
The tools are delivering. At some point, you will have to decide what happens to the surplus, or let the default decide for you.
What does your organization’s reinvestment decision actually look like right now? I’d genuinely like to know. Drop it in the comments.
p.s. If you’re leading an AI transformation and want a framework for moving from Harvesting or Experimenting to Building, the AI Lead Accelerator is built for exactly that: hyperadaptive.solutions/accelerate.
And if you want the full roadmap — five stages, case studies, and the frameworks we referenced above — Hyperadaptive: Rewiring the Enterprise to Become AI-Native releases May 12th. Pre-order at hyperadaptive.solutions/book.