How an engagement works

From a five-minute audit to AI your team can actually trust.

One deliberate ladder. Each rung stands on its own — you only climb it if the last one earned it. No funnel, no lock-in, no black boxes.

The ladder

Audit, pilot, retainer. In that order.

01Start here · free

The Audit

A five-minute self-serve scorecard, or a deep evidence-based audit. It shows where your UX and AI maturity actually stand — and surfaces the specific workflows where AI could help, and where it needs a human gate.

Free to start · deep audit £4,500 · 10 days

You leave with: A clear, honest picture and a shortlist of workflows worth acting on.

02The build

The Draft-and-Approve Pilot

We take one surfaced workflow and design the human-approval layer for it: the agent drafts the action, your person approves it in a glance, and nothing acts unattended. Built on top of the systems you already run — we design the layer, never the plumbing.

from ~£6,500 · ~4–6 weeks · one workflow

You leave with: A working approval interface, an audit trail, and a handover pack you own outright.

03Keep + grow

Support & new workflows

Every pilot includes two to three months of check-in support — we make sure it’s working in the real world, at no charge. After that, ongoing care is optional and only if you want it. The real paid work is new workflows, scoped as their own mini-pilots. We’d rather earn the next piece than bill you for a subscription you don’t need.

Check-ins included · care optional · new workflows priced per scope

You leave with: A layer that keeps working — and a partner who charges for value, not for standing still.

Support sits on the pilot, not the audit — there’s no layer to care for until one is built. The included check-ins come with every pilot; anything ongoing is a choice we’ll talk through, never spring on you.

Every rung and price, on one page →

The other door · For SMEs adopting AI

The Business Diagnostic — know where AI actually helps, before you spend on it.

You run an operationally complex business — enquiries, quotes, scheduling, suppliers. You know AI and better structure could save real time somewhere; the question is where, and what it’s worth. The diagnostic walks your enquiry-to-quote-to-delivery journey, finds where the hours and margin leak, and tells you plainly what to fix, what to automate, and what to leave exactly as it is.

£950 fixed · delivered in days, not weeks

You get:

  • A findings report on your enquiry, quote and operations journey.
  • Prioritised fixes — what to change first, and what it’s worth in time and money.
  • A clear what-not-to-automate list, so the service your clients value stays human.
  • A recorded walkthrough of all of it, plus a follow-up by email.

Everything lands async — no workshops, no meetings unless you want one. It’s also step 2 of the Funded AI Adoption sequence — the same five lenses as the WFK audit, pointed at your operations.

Book a diagnostic →

First champion diagnostics are running now — findings on the evidence page once clients have seen the value.

The pay-off

The point isn’t a productivity stat. It’s the design solution at work.

Yes, the numbers move — speed kept, mistakes removed, an audit trail gained. But the real pay-off is watching the design do its job: the same AI, now accountable. Every action carries a name, every decision leaves a record, and the person in the loop gets back control they can exercise in seconds — not a dashboard they’ll ignore. That’s the difference between AI a team quietly abandons and AI that becomes infrastructure.

Why now · Accessibility

Accessibility just became an operating requirement — and AI is quietly breaking it.

The European Accessibility Act has been enforceable across the EU since 2025, and the AI features you shipped this year are digital products in scope. But most accessibility auditors are equipped for static pages — not for an agent that streams its output, acts on someone’s behalf, and hands control back to a human.

Worse, AI code generation is measurably regressing accessibility — the surfaces you ship fastest are often the least accessible you own. So this isn’t a checklist you pass once. It’s a capability you operate: checked continuously, gated by a human, and logged — the same approval-and-audit layer we build, pointed at compliance. The record becomes your evidence.

The audit already tests it — against the four questions a page scan can’t reach: can every user perceive the agent’s reasoning, operate the approval gate, understand the action, and recover when it’s wrong?

See where you stand →

Indicative findings and recovery design — not a formal conformance assessment. Formal certification is referred to a certified partner.

Straight about the boundary

What we do, and what we don’t.

We do

  • Design the approval layer — the review surface, the gate, the record.
  • Hand you something you own outright, on the systems you already run.
  • Keep a human in the loop at every consequential step.

We don’t

  • Rebuild or integrate your systems, or own your data plumbing.
  • Let an agent act unattended — that’s the whole point.
  • Lock you in. You can take the handover and run it yourself.
Start the conversation

Begin with the audit. It’s free, and it’s genuinely useful on its own.

No pitch deck, no discovery-call funnel. See where you stand in five minutes — everything after that is a choice you make, not one we assume.