Two engagements. Both fixed-price, both scoped before you commit.

One ranks your candidates. One builds the winner. There's no third thing, no discovery phase, and no open-ended retainer.

Start here if you have four candidate problems and no ranking

AI Opportunity Audit

$950 USD — fixed

8 business days from the intake call

You know something should change. You can't rank the candidates, and guessing wrong costs more than the audit does.

What you get

  • A 45-minute intake call
  • A review of up to three existing workflows, as you describe them
  • A ranked, costed shortlist of three automation candidates — each with roughly what it costs you in time today, roughly what it costs to build, and the reason it is or isn't first
  • A one-page build spec for the #1 candidate: the trigger, the steps, the tools, who owns it afterwards, and what "working" means in testable terms
  • A 60-minute walkthrough call to take you through it

What it is not

  • No access to your systems — this works from how you describe them
  • No build, no configuration, no vendor negotiation
  • No written strategy beyond the shortlist and the one-pager

The part that matters

The $950 is credited in full against a build if you start within 60 days. Proceed, and the diagnostic was free. Don't proceed, and you still keep a ranked shortlist you can hand to anyone — including your own team.

Start here if you already know which workflow is broken

First Workflow Build

$4,500 USD — fixed

5 weeks from the signed spec

One workflow, running in daily use, on the tools you already pay for. Not a prototype, not a demo — the thing your team uses on Monday.

What you get

  • One workflow built and running in daily use on your stack — your CRM, your email, automation on n8n or equivalent
  • All configuration done
  • A written runbook, including the failure modes and what to do about each one
  • A 45-minute handover session with the person who'll own it
  • Two weeks of hypercare — anything we built that breaks, we fix, free

The scope bound — read this before you ask us to quote

At most three systems touched, at most ten workflow steps, one trigger. That bound is the reason five weeks is a real number instead of an optimistic one.

  • A second workflow — that's a second build, and we won't start it until the first is running
  • Custom software, model training or fine-tuning, data migration
  • Tool licences and API usage — your own accounts, billed direct. We never resell or front tool costs
  • Anything outside the spec you accepted. A change request becomes a new build, quoted separately

Why fixed price, and why no guarantee

The real fear with a first engagement isn't the price, it's the open-ended bill. So the price is fixed against a spec you accept before we start any work, and the scope bound above is written into it. We don't offer a money-back guarantee — we're a new practice, and a refund promise we might not be able to honour is worth less to you than a scope we can actually hold. Fixed scope plus two weeks of hypercare does the same job honestly.

Founding Build — $2,950 for the first three

The first three builds go out at $2,950 instead of $4,500 — a real $1,550 off, not a struck-through anchor. In exchange we ask for written permission to publish the workflow we built (named or anonymised, your call), a short quote, and a 30-minute conversation after handover.

We're straightforward about why this exists: we're buying proof, not volume. We have no published work, and the only honest way to get it is to build something real for someone who'll let us show it. You get $1,550 off and direct access to whoever is building it.

Closes at three builds sold, or 31 March 2027, whichever comes first. Never extended and never renewed — if we can't sell three by then, the problem isn't the price and another discount isn't the answer.

Founding rate — the first three builds

What we'll turn down

Declining is most of what keeps a small practice honest, so here it is in writing rather than discovered on a call.

We don't build or train models, and we don't touch regulated data

No custom ML, no data-science pipelines, no model hosting — we compose hosted models and automation tools that already exist. And we don't take workloads whose core is protected health information, cardholder data, or privileged discovery. We can't offer the security posture or the audit trail those deserve, and pretending otherwise would be the fastest way to end this practice. If that's your core workload, you need a firm with a compliance function. We're not it — though there may be something worth doing on the non-regulated side of the same workflow.

We don't sell strategy without shipping, and we're not extra hands

No deck-only engagements. No paid readiness assessment — ours is free, which is rather the point. No discovery phase that ends in recommendations. And no open-ended monthly retainer where we're an unspecified pair of hands. Every engagement ends with something running that wasn't running before. If you want a strategy document, we're the wrong shop — and honestly, the free assessment plus the $950 audit gets you most of the way there.

We're not the fastest option, and sometimes we're not the right one

If you already have a written spec and a deadline next month, an automation freelancer or a Make/Zapier partner will do it cheaper and sooner, and we'll tell you so. We're worth paying when the hard part is deciding what to build, not building it. And if the honest answer is "buy the £30-a-seat tool and change nothing else", that's what we'll say — it costs us the sale and it's still the right advice.

Not sure which one you need?

That's what the assessment is for. It scores which of the two is right — or tells you that neither is, yet.

Take the free assessment