Tempest

Yesterday I spent most of the day trying to get the first founders into the alpha of something we’re building.

We got 0 applications.

At first that obviously doesn’t feel great, but it’s also useful information. It forced me to question whether I was asking people to try a product before properly understanding whether they actually experience the problem.

We’re building Tempest, a decision workspace for founders navigating fragmented context.

The problem we’re exploring is what happens when the context behind a startup gets spread across customer conversations, AI chats, notes, docs and your own head. You make a product decision, move on, and a few weeks later you remember what you chose but not always the full reasoning behind it.

So I’m changing how I approach the alpha.

Instead of starting with “want to try our product?”, I’d rather understand how founders currently deal with this.

If you’re actively building something, how do you keep track of the reasoning behind important product decisions, especially when customer feedback, AI conversations and your own assumptions all played a role?

You probably remember what you decided.

But can you reconstruct why?

A customer said something in a call.

You explored it with ChatGPT.

There was an assumption in a doc somewhere.

You considered three directions.

You rejected two.

Then you shipped one.

Three weeks later, you remember the decision, but the reasoning behind it is scattered across chats, notes, docs, and your own memory.

I've been thinking a lot about this while talking to founders.

Tasks are usually tracked.

Decisions sometimes are.

But the evidence, assumptions and reasoning behind those decisions often aren't.

Curious how other founders handle this:

When you need to revisit an important product decision from a few weeks or months ago, where do you actually look?

We kept noticing the same problem while building: context ends up scattered across AI chats, notes, docs, customer conversations, and our own heads.

You remember what you decided, but a few weeks later the evidence, assumptions, and reasoning behind it are much harder to reconstruct.

That’s the problem we’re building Tempest around.

Curious how other founders handle this today. Do you have a system for keeping decision context connected, or is it mostly scattered?

We’re opening the alpha now:
https://tempest.clarantica.com