The starting point · Service 01
Get more out of what you already pay for.
Most firms are already paying for more AI than they use. A Microsoft 365 or Google Workspace subscription, a couple of ChatGPT or Claude seats, and almost none of the capability switched on. This is the work of finding what is already there, turning it into something the firm can actually use on a Tuesday, and teaching the people who will use it. Nothing gets built. Nothing new gets bought. It runs on the flat per-seat fee you are already paying rather than a metered bill that grows with use.
Who it’s for
Firms that tried AI, got a confident answer that was wrong, and quietly concluded it does not work for their kind of work. It usually does. It just arrived knowing nothing about your firm, with none of your templates, none of your standards and nobody to show anyone where it can be trusted. That is a setup problem, not a capability problem, and it is the cheapest one here to fix. It is also the right first step for anyone who is not yet sure whether they need something built at all.
How it runs
01
A conversation
Half an hour, free, and no preparation needed. It usually ends with me telling you which of these three you actually need, which is sometimes none of them.
02
Half a day with you
Going through what the firm does all day, what it already subscribes to, and which of that is switched off. Your people, not a survey.
03
A written recommendation
What to turn on, what to leave alone, what should never go near a tool that is not yours, and whether a subscription is genuinely enough for you.
04
The setting up
A week or so, depending on scope. Skills, templates, saved prompts and connectors configured inside the accounts your staff already sign in to.
05
A session with the people who will use it
On their own work rather than a demo, then written up and left with you, so it survives whoever learned it moving on.
What gets worked out first
Half a day going through what your firm does and what it already has. This is the part that decides whether you need anything else at all, and it is the part I would rather you paid for than skipped.
- What you already subscribe to, and the capability nobody has switched on
- Which everyday tasks are genuinely worth handing over, and which are not
- Where a model will confidently get it wrong in your particular field
- Whether a subscription is enough, or whether something actually needs building
What gets set up
Configured inside the tools your staff already have open, so there is no migration, no second subscription and nothing to unlearn. All of it stays yours if you later change providers.
- Skills, projects and instruction files carrying your templates and house standards
- Connectors pointed at what you already run: your file share, your inbox, your job records
- The right model chosen per task, so routine work is not run on the most expensive one
- Saved prompts your team writes once and everybody reuses
- Checks on the work that matters, so a wrong answer gets caught rather than filed
What your staff actually learn
Sat down with the people who will use it, on their own work, not a demo. The tools are the easy part. Knowing when to trust the answer is the job.
- How to ask for something in a way that gets the answer you wanted
- Which tasks to hand over and which to keep, with reasons
- How to tell a right answer from a confident one, in your own subject
- What must never be pasted into a tool that is not yours
- Written up and left behind, so it survives the person who learned it leaving