How we work
AI projects usually go wrong in predictable ways: no clear goal, the wrong tool, data nobody can get to, or staff who were handed a login and nothing else. The way we work is built around avoiding those.
Four steps, each with a clear result
The same structure applies to a two-week assessment and a three-month build, so you always know what stage things are at.
01Discover
We talk to the people who do the work, watch how the key processes really run, and look at the systems and data involved. On a project, this takes days. In an assessment, it's most of the engagement.
You get: An agreed picture of the problem, and an honest view of whether AI is the right fix for it.
02Propose
You get a written proposal: what we'll build or configure, what it costs, how long it takes, how we'll know it worked, and what we need from your team.
You get: A fixed scope and price before any work starts.
03Build
We start with a manageable piece of repetitive work, so you see something working within the first couple of weeks. We review progress with you weekly and test against real examples from your business before anything goes live.
You get: Working, tested systems running in accounts you own.
04Hand over
We train the people who'll use and manage it, write documentation a non-specialist can follow, and stay close for the first weeks of real use while things settle.
You get: A team that can run it without us, and the option of ongoing support if you want it.
Ways to work with us
Most clients start with an assessment or a single project, and some move to ongoing support afterwards.
| Assessment | Project | Ongoing support | |
|---|---|---|---|
| Best for | Working out where AI fits and what to do first | A specific rollout, automation, or tool | Keeping things running and improving |
| Typical length | 2–3 weeks | 2–12 weeks | Month to month |
| Pricing | Fixed fee | Fixed price, quoted after scoping | Flat monthly fee |
| What you get | Prioritised roadmap and a walkthrough together | A working, documented system and trained staff | Monitoring, updates, monthly summary, quarterly planning |
| Your time | Conversations and a review of recommendations | A weekly check-in and some testing | A monthly review |
How we handle your data and systems
Connecting AI to a business means giving software access to sensitive information. These are the rules we work to.
Business-grade plans only
We set up AI tools on business or enterprise plans whose terms exclude your data from model training, and we check the retention and sharing settings ourselves.
Minimum access
We ask only for the access a task needs, work inside accounts you own, and hand back or remove our access when the work is done.
Data flows written down
Every integration comes with a plain-language record of what data goes where, which services process it, and who can see it.
A person signs off
Where AI output reaches customers, money, or anything regulated, we build in a review step so a person approves it first.
Confidentiality
We'll sign your NDA before discovery. We never mention clients or their work publicly without written permission.
You own the work
Code, automations, configurations, prompts, and documentation built for you are yours, and keep running if we part ways.
What you can hold us to
We'll tell you when AI isn't the answer
Even when that means a smaller project, or none.
Plain language
You'll understand what we recommend and why, without needing to learn the jargon.
No surprise invoices
Scope and price are agreed in writing. If something changes, we discuss it and you approve it before any extra cost.
No commissions
We don't take referral payments from software vendors, so a recommendation from us is based only on what fits your business.
Start with a free call.
Thirty minutes, no obligation. We'll listen, ask questions, and tell you honestly how we'd approach it.