AI Transformation & Automation

This is for you if

• Your team is buried in repetitive work that "could probably be automated," but nobody knows where to start.

• You bought AI tools before and nothing changed, because they never fitted how the work gets done.

• You want AI inside your workflow with clear control points, and someone accountable for where it goes wrong.

• You want to know what automation will cost to run before you commit to it.

What we do together

• Map your processes and find the bottlenecks — where time leaks and how much.

• Sort every task into three tiers: fully autonomous, AI behind a human review gate, humans only.

• Cost each candidate: tokens, infrastructure and tooling per month, against the hours it returns and the payback period.

• Set up the infrastructure for your needs — hardware, software, models, context, skills, fine-tuning where it pays for itself.

• Bring your team onto the new way of working, with review steps that hold quality as the pace goes up.

How it works

1. Discovery calls, free. Up to three: two to understand your context and look over what you already have, a third where I come back with a proposal and a price.

2. Process audit. €1,500 for a single team or process area, €3,000 where the scope is wider. The map, the tiering, the costings and a milestone plan, written to be taken into a management meeting and signed off.

3. Pilot on one process. Priced whole against milestones once the audit has set the scope. Fast result, low risk, real numbers to check the audit against.

4. Scale and train. Project fee, or a retainer from €2,000 a month covering a named area of ownership. Up to two days a week of focused work, plus availability within one working day for anything urgent. Minimum three months, monthly after that with 30 days' notice on either side.

What you get

A process map with every task tiered as autonomous, review-gated or human-only. Each candidate costed per month against the hours it returns, with a payback estimate.

Working automations for the tiers that moved, running in your own infrastructure.

A team that uses them, and the review steps that keep quality where it was.

A prioritised list of what to automate next, with the recommendations behind it. You can run that yourselves, or keep me on to review and keep automating as the processes change.

Proof

Moved a development team onto AI agents and rebuilt the process around it. Typical feature delivery went from two days to two hours, review discipline held, and technical debt was cleared.

Replaced an hourly-paid moderator with an automated pipeline that ran during projects instead of after them, on a platform that has run over 900 studies. Data quality improved roughly 5–7x, because respondents whose recording conditions defeated computer vision stopped flattening the results.

Built automated client report generation in early 2022, before generative AI hyped, by scripting a slide template against the study database.

Set up Label Studio and trained four non-technical colleagues to label training data.

My own automations currently save 2–3 working days a week.