Numbers an owner can trust, from someone who knows when numbers lie.

Guidepost exists because of a gap that bothered me for years: big companies pay analysts to read their numbers and flag what matters. A ten-truck HVAC shop generates just as much data (in Jobber or Housecall Pro, in QuickBooks, across ad platforms) and gets nobody. The owner is the analyst, at 9pm, in a spreadsheet.

So Guidepost does that job: it reads the numbers you already have and sends you a short, plain-English note about the few things that need attention (money you’re owed, marketing that’s slipping, quality issues forming), with every figure traced back to its source.

Why the restraint matters

Guidepost is built by a small team led by a founder with a background in data science: the discipline of knowing when a number proves something and when it only hints. That distinction is the whole product. Every AI tool will happily hand you a confident answer; the hard part is knowing which answers deserve confidence. So Guidepost draws a hard line:

• When a fact has an obvious next step (overdue invoices, a cluster of callbacks), it says so plainly.
• When acting would mean predicting the future (like moving ad budget), it shows you what changed and leaves the call to you.
• Every number shows its work. The figure, the definition in words you’d use, and the source. If the data is too thin to answer, it says “not sure yet” instead of guessing.

Where it's at

Guidepost is in early access. I’m onboarding the first handful of home-service and trades shops personally (setup included) in exchange for honest feedback. If that sounds like a fair trade, get in line or just email me: hello@getguidepost.ai. I read every reply.