Our method, published blank
The templates we fill in when you hire us, empty and free. Nobody buys a template — so you can have the format, and judge from it whether the paid version is worth $950.
What's here
Ranked shortlist template
The exact format the $950 audit delivers: three automation candidates, each with the time it costs you today, the rough build cost, and the reason it is or isn't first. Download the blank template and see exactly what you'd get.
"We've been experimenting with AI and nothing has changed"
How to diagnose why AI experiments fail: the four signals (clear bottleneck, system of record, named owner, authority to approve) and what's missing. If you've tried AI for months with no results, this tells you what to fix first.
Workflow teardown: Inbound lead qualification
Reference design for inbound lead qualification: trigger, steps, failure modes, build cost ($3,000-$6,500), and ongoing cost (~$1,100/month by hand). Labelled as reference design, not client work.
More resources coming
We're adding new teardowns and templates on a biweekly cadence. Each one demonstrates a capability instead of claiming one — because a new practice has no case studies, only shown work.
Check back for: new-client intake teardown, proposal first drafts teardown, overdue invoice chasing teardown, weekly status reporting teardown.
Frequently asked questions
Is the assessment really free?
Yes — the assessment and the written diagnosis cost nothing, and there's no call required to get them. There is no paid version of the assessment. If you've seen a $2,500 "Foundation Assessment" on this site before, that was an error and it's gone; the two paid engagements are the $950 audit and the $4,500 build.
You're new. Why should I trust you with this?
You shouldn't, on our word. We have no published engagements — we're a new practice and we're not going to invent any. So we've published the method instead: the templates on this page, the workflow teardowns, and the assessment itself, which is built to tell roughly six in ten firms they're not ready. Read the teardowns and judge the thinking. If it doesn't hold up, you've lost twenty minutes and learned something about what a build involves. That's a harder test than a logo wall, and it's the one we'd rather be set.
What if the build turns out to be bigger than the spec?
Then it's a second engagement, quoted separately — and you'll hear that during scoping, not in week four. The bound is at most three systems, at most ten steps, one trigger. If your workflow doesn't fit inside that, the honest answer is that it's either two builds or the wrong first project, and we'd say so before you paid anything.
How fast can you start?
Often not as fast as you'd like. There's one build in flight at a time. If the slot is full, the assessment shows you the next real start date instead of a booking link. If you need something live inside a month, we're the wrong choice and an automation freelancer is the right one — we'll tell you that rather than take the work.
Which tools do you use?
Yours, wherever possible — your CRM, your email, whatever you already pay for. We compose hosted models and automation tooling (n8n or equivalent) rather than building software. Licences and API usage stay on your own accounts, billed direct to you — we don't resell tools or mark them up, so there's nothing we're incentivised to recommend.
Can you build us a custom AI model?
No. No custom ML, no fine-tuning, no model hosting, and nothing whose core is regulated data — patient records, cardholder data, privileged discovery. That needs a firm with a compliance function and we're not it. What we do is compose tools that already exist into a workflow that runs.