AI consulting for businesses usually starts at the same point: people are already using AI tools at work, often on personal accounts, and someone now needs to put it in order. I help UK teams choose the right AI tools, set them up properly on business accounts, and write a usage policy people will actually follow.
The work is practical. When it is finished, your staff know which tools they may use, what they must never paste into them and who to ask when they are unsure. You know what you are paying for, who controls each account and where your data goes.
What changes when this is done
A common starting point looks like this: three or four people paying for different chat assistants on their own cards, client details pasted into free versions, nobody checking the output before it reaches a customer, and no written rule about any of it. The business is already using generative AI; it is just doing so without anyone owning the decisions.
Afterwards you have one or two approved tools on accounts the business controls, sensible data settings switched on, a short written policy and a named person responsible for it. Staff who were nervous about using AI have clear permission for the tasks it suits. Staff who were using it carelessly have clear limits.
Who this suits, and who it does not
It suits teams of roughly three to fifty people in service businesses, practices, agencies, retailers and charities, where AI is being used for writing, summarising, research, admin and customer replies. It is especially useful if you handle client personal data, or if a client, tender or insurer has asked to see how you control AI use.
It is the wrong fit in three cases. A sole trader who simply wants to get better at one assistant does not need a policy or an admin console. An organisation using AI to make significant decisions about people, such as credit or recruitment screening, needs a solicitor and a data protection specialist first. And if the real question is what marketing to do at all, start with a written digital marketing strategy rather than tools.
What the engagement covers
A look at how AI is used now
I start with a short, honest questionnaire for staff: which tools they use, for which tasks, on which accounts, and what kind of information goes in. Alongside it I review the subscriptions on company cards and expense claims. People answer more truthfully when it is clear the aim is to make their use safe, not to catch anyone out.
Choosing the tools
I match the tasks your team actually does to the tools that fit them. For most small businesses that means one general assistant built on a large language model, such as ChatGPT, Claude, Microsoft Copilot or Gemini in Google Workspace, plus perhaps one specialist tool for transcription or images. The choice depends on:
- whether you already work in Microsoft 365 or Google Workspace;
- what the business plan’s terms say about using your data to train models, and how long conversations are kept;
- which admin controls are available, such as single sign-on and removing a leaver’s access in one step;
- where the provider processes data, and what that means for transfers under UK GDPR;
- the real monthly cost per seat in pounds, bearing in mind that several tools are priced in US dollars.
Shortlisted tools are tried on samples of your own work before anyone commits to annual licences. The software I use in my own work is listed on my tools page, but your shortlist is decided by your tasks, not my habits.
Setting the tools up
I set up the business workspace, add users from company email addresses, and switch on the controls the plan offers: training on your data turned off where the plan allows it, retention set to something you can justify, and connections to mailboxes and shared drives enabled only where they are needed. Where it helps, I write shared instructions, sometimes called a system prompt, so the assistant knows your tone, your services and what it should never claim. Then I help move people off personal accounts so that business conversations stay with the business.
A written AI usage policy
The policy is a plain document of a few pages, not a legal text nobody opens. It covers which tools are approved and how to request a new one; which information never goes into any AI tool, including special category data such as health details, passwords and client material under confidentiality terms; how outputs are checked before they are used; when you tell clients AI was involved; and who reviews the policy and when.
It follows the Information Commissioner’s Office guidance on AI and data protection and the principles of UK GDPR. I also help you work out whether each provider acts as your data processor, whether a data processing agreement is in place, and whether any planned use is risky enough to need a data protection impact assessment.
Working guides for each role
Each role that uses the tools gets a short guide: what AI is good for in that job, what it is not, and a handful of tested prompts for the tasks they repeat every week. These are stored where the team already keeps documents, so they get updated rather than forgotten.
What I will not do
- Give legal advice. The policy is a practical working document. Where you handle sensitive data or work in a regulated sector, have your solicitor or data protection officer review it.
- Promise time savings. How much time AI saves depends on the tasks and the people, and any figure quoted before you start would be a guess.
- Build chatbots, agents or automations under this engagement. Those are separate pieces of work, scoped once the basics are in place.
- Recommend a tool on reputation alone. Everything on the shortlist is tried against your own tasks first.
How an engagement runs
- A free call to understand your team, your current tools and what is worrying you.
- The current-use review Staff questionnaire, subscription check and a conversation with whoever handles data protection.
- Shortlist and trial Two or three people test the shortlisted tools on real tasks, and we compare results side by side.
- Setup Accounts, users, data settings, connections and shared instructions.
- Policy I draft it, we go through it together, and you sign it off.
- Handover A session with the people who will use the tools, plus the role guides.
- Follow-up Whether there is a later review of what is used, what is not and what the policy missed, and when, is agreed before work starts.
Problems I see most often in UK teams
- Business history in personal accounts. When someone leaves, months of client conversations leave with them, still sitting in an account the business cannot reach.
- Free versions used for client work. Consumer tiers can have different data terms from business plans, and staff rarely check which they are on.
- A template policy that bans everything. It gets ignored within a week, which is worse than having a realistic one.
- Nobody checks the output. AI tools state wrong facts confidently, a problem known as hallucination, so a named person needs to review anything that goes to a customer. That review step is what people mean by human in the loop.
- Duplicate subscriptions. The same business pays for two or three assistants that do the same job.
- Mass-produced web content. Publishing AI drafts unedited can hurt search visibility; there is a separate guide on using AI for marketing without hurting SEO.
What you receive
- A short report on current AI use and spend, with what to keep, change or stop.
- A tool recommendation with the reasons, costs in pounds and data terms for each.
- Configured business accounts that you own, with your administrator holding control.
- A written AI usage policy, ready to share with staff and, where asked, with clients.
- Role guides and a starter set of tested prompts.
- A record of the data protection checks made, to keep with your other UK GDPR documents.
What it costs
The price depends mainly on team size and the number of tools involved, and it is agreed in writing before work starts. Each quote is written in GBP after a free first call. Tool subscriptions are billed to you directly by each provider, so you never pay for them through me.
Next step
Book a free 30-minute call and tell me which tools your team uses now and what you would like to be true in three months. I will tell you whether this engagement is the right fit or whether a lighter fix would do.