This week, roughly $285 billion was wiped off enterprise software valuations. The trigger: AI agents that handle contract review, compliance, sales preparation, research — work that previously justified thousands of SaaS seats.

The market's conclusion: AI kills software.

We've been here before.

SaaS was going to kill on-prem. It didn't. What it actually did was create a decade of hybrid complexity, integration debt, and data sovereignty challenges most enterprises are still untangling. Capital panicked then, too. The pendulum swung. Then it grounded.

This pendulum will ground as well. Carriers won't abandon decades of policy administration infrastructure. Banks won't rip out core platforms. The $285 billion will return — selectively — to vendors that demonstrate something more convincing than a press release.

But AI does create a genuinely new risk. Just not the one capital is pricing.

I wrote recently about the monolithic noose and death by a thousand cuts. The noose is legacy — everyone knows it. The thousand cuts is what happens when AI adoption runs without architectural governance. A thousand intelligent workflows nobody can standardise, maintain, or scale. Fragmentation disguised as progress.

Most vendors are heading straight for the thousand cuts. Adding copilots. Adding assistants. Branding everything "AI-powered." The Hartford's CEO declared the company "AI-first" — "the sky is the limit." Bold. But AI layered onto existing workflows is a powerful multiplier applied to a single dimension. Efficiency gains that never compound into anything structurally different.

The vendors that succeed won't be the ones that bolted it on fastest. They'll be the ones whose product leaders understood that AI changes what software fundamentally is — not what it can do, but how it learns, how it compounds value, how it evolves with every interaction.

That's a design problem. Not a technology problem. Intelligence woven into the product's principles rather than sprayed across its surfaces.

Most software organisations aren't having that conversation. They're having the earnings call version — efficiency metrics, cost reduction, percentage improvements. But the moment a competitor designs AI into the foundation rather than the interface, bolt-on vendors discover they've been building on sand.

Software vendors need to come out swinging. But not with bolted-on features and "AI-powered" press releases. They need to be hiring product leaders who think beyond the knee-jerk — leaders who understand this is an architectural redesign, not a technology upgrade.

Do you want your next earnings call to be "we've bolted AI onto the side of our platform" — or do you want to be laying foundations for long, stable financial success?

The slides walk through the difference. And if it's the latter — I'm easy to find.

#AI #EnterpriseTransformation #InsurTech #ProductLeadership #SaaS #CPO

First Comment

The SaaS parallel is worth sitting with.

In 2010, capital said on-prem was dead. By 2015, every enterprise was managing hybrid infrastructure nobody had planned for. Cloud-first became cloud-also became cloud-and-everything-else. The vendors that won weren't the ones who moved fastest — they were the ones whose architecture absorbed the shift without breaking.

AI is the same pattern at higher stakes. The vendors branding themselves "AI-powered" today are in the same position as the ones who stamped "cloud-ready" on unchanged products fifteen years ago. It bought time on the earnings call. It didn't buy a future.

What buys a future is designing AI into the way your product thinks — not just what it does. Software that learns from every interaction. Intelligence that compounds over time. A product that gets better with use, where switching costs are accumulated knowledge rather than contractual lock-in.

That's a fundamentally different product. It requires a fundamentally different kind of product leader.

More on this soon. I'm easy to find.