#57 When a Chairman Builds Software, the Moat Is Shrinking

I was a few minutes early. Martin Grimberg, the Vizcon chairman, was already at the table. Before I'd even sat down, he turned his laptop toward me.

– I'm building a board management tool, he said.

I laughed. Three days earlier, I'd done exactly the same thing.

Since beginning to step back from the CEO role at Ambition, I've been thinking more practically about what board work actually requires. Four boards, more focused involvement, more administration… 

I'd looked at the available tools provided by different software companies over the years. Used a few of them. All felt slow, overbuilt, and slightly beside the point. So I wrote an extensive brief in Claude, trying to capture most of my thoughts on board work, and then took it into Lovable. By the end of the afternoon I had a working prototype. Not a product, proper engineering would still be needed for that, but possibly something good enough for my own purposes.

Thirty years ago, Excel did something similar. Suddenly anyone could build their own mini-application. Rough, custom, good enough, and often better than the packaged alternatives. Not production software, but close enough for most purposes. Millions of those spreadsheet apps are still running today, somehow holding businesses together. Nobody planned them. Someone just had a problem and a few hours. Nobody wrote documentation either. Nobody needed to. At least not until the person who built the pricing model retired, or moved on.

AI generated tools are a second wave of the same, but with greater possibilities and bigger risks. The tools are more capable, but the output is harder to read. The chairman prototyping his own board tool on a Tuesday evening is the same person who built the pricing model in Excel in 1996.

And when a chairman can build his own tool in a few evenings, that creates a real problem for a lot of software companies.

Some of these companies have spent tens of millions building what they have. That investment sits on the balance sheet as an intangible asset, the software itself priced as though it was expensive to build and expensive to replicate. That is not true anymore, at least not across the board. The auditors haven't fully caught up yet. But they will.

A lot of software goodwill is going to be written down. Not all of it, though. The software that coordinates hospital logistics or manages financial risk has genuine worth. The more embedded in critical operations, the safer the valuation. But the part of the market that is more like a glorified spreadsheet application with a decent UI and a sales team is exposed. The moat was the development cost, and the development cost is now greatly reduced. 

So what replaces this moat? Probably no one has fully worked this out yet, including the companies most affected. Data that accumulates and compounds the more people use the product is part of it. Workflow complexity so embedded in critical operations that replacing it would be more painful than living with it. And the hardest to replicate. Knowing how a specific kind of work actually gets done. Not the documented version. The real version. That knowledge comes from being there, and you can't shortcut it with a synthetic persona.

My board management tool is still sitting in Lovable, unfinished. I'm guessing Martin's is too. Maybe I'll finish it, maybe I won't. But the moat question feels more urgent every time I open that prototype. Software is a weaker moat than it has ever been.