80% of the value of Clay.
1% of the time.

A founder letter from Collin Stewart.

I’m a huge Clay fan.

It’s the most powerful sales tool ever built. If you know how to use it, you can do almost anything: find any company, enrich it with anything, score it however you want, route it wherever you need. It’s a superpower.

But I got tired of clicking buttons.

I spent the last few years running outbound for a portfolio of startups — building Clay sheets, configuring agents, wiring up enrichment chains, hand-tuning columns until the data looked the way the founder needed it to look. And I noticed two things.

One: most of what I was building was structurally similar. Same pipelines. Same enrichment patterns. Same scoring loops. Just retuned for each customer. I was rebuilding the same machine over and over with slightly different paint.

Two: I had become a translation layer.

Salespeople and founders would come to me with brilliant ideas — “find every company that just hired a Director of Operations in cities where commercial rents jumped 10% YoY” — and they couldn’t execute on them. Not because the ideas were hard. Because the tool was hard. Clay’s power lives behind a steep curve. You have to know when to use a Claygent and when to use a Clay credit. You have to know how to write a Claygent prompt that actually finds what you asked for instead of hallucinating something close enough — or burning hundreds of dollars on the wrong sample size before lunch. You have to know how to push an array out of a JSON object into another sheet, and how to reverse-lookup to join two tables when you forgot to do it the first time. They had the ideas. I had the muscle memory. They got bottlenecked on me.

That bothered me. Everybody should be able to use Clay. But you can’t. So I built something else.

Predictable.ai is the version of Clay where the muscle memory belongs to the AI.

You describe your ideal customer in plain English. The system goes and finds them, evaluates them against your criteria, ranks them by tier, finds the right people inside, validates their emails, and watches for buying signals — funding rounds, new hires, job postings, news mentions. No sheets. No webhooks. No conditional logic to debug at midnight.

Under the hood, it uses many of the same ideas you already love about Clay. Waterfall enrichments across providers — ai-ark for contact data, ZeroBounce for email validation, and a handful of others — so you stop thinking about API keys and which tool returns the cleanest data. Source companies with whatever weird characteristics you want. Find the people inside. Run them through a classifier with your favorite AI to filter down to the real fits. The real star of the show, though, is Exa Websets. That’s the primitive that lets us cover 80% of what Clay does without you ever building a sheet.

Then we added two things Clay doesn’t do natively. Yes, contacts flow straight into Instantly (and your CRM) when they’re ready to send — that part’s table stakes. But the wizard also drafts the sequences for you, using the Claude Code skills I’d already built for my own clients. So it’s not just accounts and contacts. You get accounts, contacts, and the outbound campaign — written. Idea to production in five minutes.

That’s the bet.
80% of the value of Clay. 1% of the time. 0% of the learning curve.

If you’re a founder trying to find your first 100 customers, and you don’t have a Collin sitting in the next office to translate your ideas into Clay sheets — get on the list. We’re letting people in steadily.

— Collin Stewart

P.S. I still love Clay. If you’re a power user with the time to build, you should keep using it. Predictable is for everyone else.