How to know you're ready for AI. Five honest signals.
Not every business is ready for custom AI. Here are the five signals that mean you actually are, and the two that mean you're not yet.
Most "AI readiness" content is vendor marketing disguised as advice. We're going to be specific, because you can't fake these signals.
If you have three or more of the five below, you're ready. If you only have one or two, you're not, and no amount of consulting theater will change that.
Signal 1 — you have a repeatable, expensive manual process
Somebody on your team does the same thing every week. Every day. Processing invoices. Writing status reports. Reviewing submissions. Answering the same 40 questions.
That repetition is the substrate of AI. It means there's pattern, there's input, there's output, and the cost of doing it manually is measurable. Put a number on the hours per week and the loaded labor cost. If it's over $50k a year, you have signal 1.
Signal 2 — you have data that nobody is using
Not "big data." Your data. The HubSpot contact records, the Intercom tickets, the Stripe invoices, the Notion docs your team wrote last year. That data is context, and context is what makes AI useful instead of generic.
If you have three years of historical operations data sitting in tools and nobody's mined it for insight, you have signal 2. An AI trained or prompted with that context performs at a completely different level than one without.
Signal 3 — your team is stretched, not broken
AI is a multiplier. If your team is already underwater drowning in broken processes, AI will magnify the chaos. You need a team that's running but stretched. Overloaded in predictable places. Good at their jobs but there aren't enough of them.
If hiring one more person would help but you can't afford to, and you know specifically what that person would do every day, you have signal 3. That's the shape of work AI replaces cleanly.
Signal 4 — you've tried Zapier or similar and hit the wall
You bought Zapier. You built six zaps. Four of them work, two of them break every month, and the thing you really want to automate involves decisions, not just moves.
If that sounds familiar, you have signal 4. Zapier and Make are amazing for deterministic work. They fail for anything that requires judgment. That's where AI steps in.
Signal 5 — you or a co-founder is technical enough to evaluate quality
AI implementations succeed when someone on the client side can look at a prompt, a response, or a piece of code and tell whether it's good. That doesn't mean they write code. It means they've been close enough to software that they can ask smart questions.
If your founder team is all sales-and-ops with nobody who can sanity-check a system, you'll end up at the mercy of whoever built it. You don't need a CTO to be ready, but you need at least one person who can press back.
Two signals that mean you're NOT ready
If your data is chaos. If your records are inconsistent, your tooling is three years out of date, and half your ops happen over text message, AI can't help yet. Clean the data first. This isn't a Foundry problem, it's a prerequisite.
If the primary driver is "the board asked what we're doing with AI." That's fine as a second reason but brutal as a first. AI initiatives led by vanity deadlines fail. AI initiatives led by specific operational pain succeed.
What to do about it
If you have three or more of the first five signals, the next step is a real audit. Ten questions, 90-second turnaround, tells you which systems to build first.
If you only have one or two, your move is different. Fix the data. Clean the stack. Hire one ops person. Come back in six months. We'll be here.