AI consulting for small businesses in Manchester
I help small and medium-sized businesses in Manchester put AI into the tools their staff already use. That means an agent in Slack or Teams that people message like a colleague: it answers questions from company documents, or fetches a figure that lives in another system. No new dashboard, no new login, no training course.
Why me
There is a lot of AI consultancy in the UK that stops at the demonstration. Someone runs a workshop and leaves a deck behind. The week after, nothing has changed. The gap between a demo and something a team uses every week is mostly plumbing.
I already build this sort of thing. At Platform81, the Stockport agency where I am Head of Paid Media, I have built connectors that let Claude query Google Ads, GA4, Meta Ads and Microsoft Ads directly, so a question about an account gets answered from live data. I also run the servers all of it sits on. Most of it is in production and used daily by the paid media team; a cross-account analytics platform I am building is not live yet. The Builds page has the detail.
None of that team are engineers. My own background is 10+ years agency-side in Manchester and an MSc in Organisational Psychology. In my experience AI projects fail on adoption rather than on the technology. A good part of the job is making the tool simple enough that people will use it.
AI automation in practice
AI agents for business processes, in Slack or Teams
The version that gets used is a bot in the channel people already sit in. Someone types a question and gets an answer from the company's own documents. Someone asks for last month's figures and the agent goes and gets them. It only has access to the systems it needs — the shared drive, the CRM — and it does one job.
Automating reporting and admin
Most small businesses still move data between systems by hand. Connecting the tools a business already pays for removes a lot of that. The monthly report and the routine data entry produce themselves, and someone checks the output. It arrives in a Slack channel or the spreadsheet the team already opens. There is a longer guide to picking the first process to automate.
The numbers are only as good as the tracking underneath them. I cover that separately, under Google Tag Manager and server-side tracking.
AI in marketing and paid media
This is the area I know best. A business running its own Google or Meta advertising already has the data; the gap is that nobody has time to read it every week. An agent connected to the account can summarise what changed and post it in the same channel as everything else. Product feed clean-up and ad copy drafted from real performance figures sit in the same category.
Built for non-technical staff
I build all of it for people who are not technical and have no wish to become technical. Nobody has to learn prompt engineering or manage another login. In practice the agent asks for what it needs and the person answers in ordinary language.
How it works
Look at how the work gets done now
Discovery is a day. I look at how the work currently gets done and produce a short list of processes, ranked by the time they take and how automatable each one is. Some come back marked as a poor fit for AI.
Build one process properly
I pick one process and build it. Narrow scope and real data, running in the client's own Slack or Teams. A pilot is scoped to weeks, not months, because it covers one process and not a programme.
Hand it over
I walk through the finished thing with whoever owns the process afterwards and hand across the documentation and all the access. I stay available for changes and for the next process. Nothing should depend on me being around.
Who it's for
A good fit
Small and medium businesses where a few people spend a chunk of every week on repetitive work. No technical team or AI strategy required. Slack or Teams plus a CRM or a shared drive is enough to work with. It helps if one person internally is interested enough to own it afterwards.
Not a good fit
Plenty of AI business consultants will write the strategy document. I would rather build one process and hand it over. A business that wants a full AI agency in Manchester, with account managers and a monthly retainer, should look elsewhere — I only have room for a few projects at a time. Same for a chatbot on the website because a competitor has one.
Common questions
What does it cost?
I quote work as a fixed piece, with the scope agreed in writing before anything starts. I price discovery separately so it can stop there. I do not work on an open-ended day rate.
How long does it take?
The gap between discovery and a working pilot depends mostly on how quickly I get access to the systems involved. Where that is straightforward it is a matter of weeks. Anything needing six months of build before anyone sees a result is the wrong shape of project for a business this size.
Do you work in person around Greater Manchester?
Yes. I am based in Manchester and can meet across Greater Manchester and the wider North West. Discovery in particular is better done in the room, watching how the work actually happens. The build itself is mostly remote.
What happens to our data?
I scope access narrowly: an agent gets the systems it needs for its task and nothing else. Where a step calls a third-party model, it runs under commercial API terms that exclude the data from training, and I will say plainly which steps send what to which provider. Where data cannot leave a company's own infrastructure, self-hosting the tooling is a real option.
Are we too small for this?
Probably not. Most of this work is taking a recurring task off two or three people, and smaller companies feel that sooner than large ones. If a process runs every week and follows a set of rules, it is worth a conversation.
Contact
AI consultant, Manchester
Describe the process in a couple of sentences and I will say whether it is worth doing. I answer my own email.