AI consultancy for UK small businesses
Practical AI consultancy for a business that needs a clear next move
A UK AI consultancy that helps you find the use cases actually worth money, set sensible rules for your team, choose tools without the hype, and plan implementation around how your business really works.
£149plus VAT for a one-hour advisory session on one defined challenge, with a written summary.
Manchester based, UK wide. Every recommendation on this page is something we run in our own business first.
The AI conversation most small businesses are actually having
It is rarely "should we use AI". It is quieter and messier than that, and it is usually already happening whether anybody decided it or not.
- A few keen staff are already pasting customer emails into free chatbots.
- Nobody has agreed what is and is not allowed, so everybody guesses.
- Confident-sounding answers get copied into client work unchecked.
- Every week brings a new must-have tool and a fresh wave of jargon.
- Somebody senior asks "what is our AI strategy" and the room goes quiet.
- The genuine time-savings sit there unclaimed because nobody was shown how.
The gap is not enthusiasm and it is not technology. It is somebody who has actually run this stuff in a real business telling you, in plain English, what is worth doing and in what order.
What an AI consultancy should actually do for you
The title has been claimed by an awful lot of people who have never shipped anything. So here is the specific work an AI consultancy owes you, and the specific things we refuse to be.
Good AI consulting is mostly translation and prioritisation. Translation, because the field moves fast and is wrapped in more jargon than any technology since the last one, and somebody needs to say what a thing actually does in terms of your Tuesday. Prioritisation, because in any small business there are a dozen places AI could plausibly help, two or three where it genuinely will, and limited money to find out which is which.
So the job is to sit down with how your business actually runs, find the points where the effort is repetitive and the input is messy, and rank them by what they would honestly return. Then to plan the first one properly: what it needs, what it risks, what it costs, and what you should see before you spend another pound. That is the whole discipline. It is not mystical and it does not need a sixty-page deck.
The other half of the job is saying no. No, that tool is a subscription looking for a problem. No, that process should not be automated because it is judgement dressed up as a rule. No, you do not need a strategy document, you need one working project and a written policy. A consultant who never says no is a salesperson with a different email signature.
What we are not
As an AI consultancy we are not a reseller. We take no commission from any tool we recommend, which matters more in AI than anywhere else right now because half the market is affiliate links wearing a lanyard. If the honest answer is a free tool used carefully, that is the answer you get.
We are not a research department either. You are not paying us to learn on your time. The advice comes from things running in our own company today, and when a question sits outside what we genuinely know, we say so and point you at somebody better placed.
Advice and implementation are different purchases
Plenty of consultants blur the two so the advice quietly becomes a proposal. We keep them separate on purpose. The advisory work is priced on its own, and it produces something you could take to any competent supplier, including not us. If you then want us to build it, that is a separate decision with a separate price, and the plan is written so you can compare.
Custom development is quoted separately for the same reason. A plan that only makes sense if you buy the build from its author is not a plan, it is a brochure.
The six things people bring us
Most engagements start with one of these. Bring yours, whatever shape it is in.
Where do we even start?
You know AI matters and that is as far as it has got. We map your workflows, find the two or three places it would genuinely pay, and put them in order. Usually this is the assessment.
AI Opportunity AssessmentStaff are already using it. Quietly.
Customer data is going into free tools and nobody agreed that. We write a short acceptable use policy your team will actually read, and train them on what good and safe use looks like.
Which tool should we buy?
You have three quotes and a free trial and they all sound the same. We review them against your actual workflow, with no commission clouding the answer, and tell you which to pick or whether to pick none.
Is this idea worth building?
You have a specific automation or AI project in mind and want it stress-tested before money moves. We map the workflow, the risks and the honest costs, and give you a plan with numbers in it.
We tried it and it flopped
A chatbot that annoyed customers, a tool nobody opened after week two. Common, fixable, and usually a scoping problem rather than a technology one. We work out what went wrong before recommending anything new.
We need someone ongoing
No appetite for a full-time hire, but you want a standing session where priorities get set, suppliers get reviewed and questions get answered by the same person each month. That is the AI Partner arrangement below.
Why we can say this with a straight face
Consulting from a business that runs this, not one that read about it
Manchester PC is a computer repair and services business. Enquiries arrive badly worded, quotes need chasing, invoices need entering, phones ring while both hands are inside a machine. It is exactly the sort of company AI consultants write case studies about, except we are the company, and we built the systems for ourselves before anybody paid us to.
Our shared inbox is triaged by AI that recognises enquiries, matches customers and drafts replies a person approves. Repair enquiries are classified against our own triage rules and quoted from a library of priced work, again with a person approving every one. Supplier invoices are read off hostile PDFs straight into the accounts. Our own receptionist line is answered by the same voice agent we sell. None of this is a demo environment. It is how the company runs on a wet Wednesday.
Just as importantly, we have the scar tissue. We automated a couple of workflows, watched them for a few weeks, and deliberately switched them back to a person because the output was not good enough for a business with a reputation to protect. We have cancelled tools that demoed beautifully and delivered nothing. When we tell you something is not worth doing, it is usually because we have already paid to find that out.
That is the actual difference between consulting from experience and consulting from content marketing. Anyone can summarise what AI might do. We can tell you what happened when we trusted it with our own customers and our own money, which parts held, and which parts we quietly took back.
Anyone can summarise what AI might do. We can tell you what happened when we trusted it with our own customers and our own money.
Read the whole storyHow an engagement actually runs
Four stages, and every one of them ends with something in your hands. You can stop at any point and keep the value.
- 01
Advisory session
One hour on one defined challenge. You leave with a written summary and recommended next actions. For a lot of businesses this is the whole engagement, and that is fine.
- 02
Opportunity assessment
A structured look across the business rather than at one question. Five prioritised opportunities, honest numbers, and a recommended first project.
How the assessment works - 03
Implementation plan
The chosen project mapped properly: workflow, requirements, solution options, risks, indicative costs and a roadmap. Written so any competent supplier could build from it.
- 04
Build, or partner
We implement it, or your people do, or a third party does. Either way the Ongoing AI Partner arrangement keeps somebody in your corner as tools and priorities shift.
What we build ourselves
Rules, training and staying on the right side of the regulator
The unglamorous half of AI adoption, and the half that stops a useful tool becoming a data protection incident. We fold it into every engagement rather than selling it as an afterthought.
Most small businesses do not need an AI ethics committee. They need a one-page acceptable use policy that says which tools are approved, what must never be pasted into them, and who checks AI-written work before it reaches a customer. Short enough that people actually read it, specific enough that "I didn't know" stops being available. We write these with you, in your words, and they take about as long as a good meeting.
Training follows the same philosophy. Not a webinar about the future of work, but hands-on sessions with your real tasks: this email, this quote, this report, done well and done safely. People leave knowing what good looks like, what sensitive data is, and why the confident tone of a chatbot is not the same thing as being right. We deliver in person across Greater Manchester or remotely anywhere in the UK.
None of what we advise is invented in-house. It tracks the published guidance from the National Cyber Security Centre on using AI tools safely and the ICO's guidance on AI and data protection. Using AI does not change your duties under UK data protection law, and any consultant who waves that away is creating your next problem.
The rule that does most of the work
AI drafts, people decide. It is on half the pages of this site because it is how we run our own company. Anything that commits money, quotes a price or goes to a customer in your name gets prepared by the machine and approved by a person. The time saved by removing that approval is minutes. The cost of one confidently wrong message sent in your name is not.
What it costs
All prices plus VAT. Advisory work is priced flat so you know the number before you book. Custom development is always quoted separately.
AI Advisory Session
£149
One defined challenge, answered properly
- A 60-minute session, in person or remote
- One defined challenge, agreed in advance
- A written summary of what we discussed
- Recommended next actions, in order
- No obligation to take any of them with us
AI Opportunity Assessment
£99 founding price
Finding where AI would genuinely pay across the business
- A structured review of your workflows
- Five prioritised opportunities
- Honest effort and cost estimates
- A recommended first project
- Full detail on its own page
AI Implementation Plan
From £495
A specific project you want planned before money moves
- Workflow mapping with the people who do the work
- Requirements written down properly
- Solution options compared, tool-agnostic
- Risks and data protection considerations
- Indicative costs and an implementation roadmap
Ongoing AI Partner
From £299/mo
A standing adviser without a full-time hire
- A monthly strategy session
- Prioritisation as things change
- Supplier and tool review, commission-free
- AI policy kept current
- Implementation guidance within the retainer
The assessment fee is a founding price while we build the case-study bank. Custom development, integrations and automation builds are scoped and quoted separately, so advisory fees never hide a sales pitch.
Advice we give surprisingly often
A consulting page ought to show you the product. This is the sort of thing an engagement actually produces, free of charge, because it is more persuasive than adjectives.
None of it is secret and none of it needs an hour to say. What you are paying for in a session is the version applied to your business, with your numbers and your workflows, and the confidence of knowing which of these rules applies to your situation and which does not.
Do not start with a chatbot on your website
It is the most visible AI project and usually the wrong first one. If your enquiries already arrive faster than you follow them up, a chatbot adds volume to a leak. Fix the follow-up first, then add capture. When chat is the right answer, it needs your real pricing and policies behind it, not a general-purpose model improvising in your name.
Your first project should be boring
The best first AI project is repetitive, internal, and checked by a person: summarising, drafting, extracting, classifying. Nobody outside the business ever sees it fail while you are learning, and the time saved is real from week one. Impressive-looking projects with customers watching are round two.
Free tools with rules beat paid tools without them
A team using a free chatbot inside a clear policy, with sensitive data kept out and outputs checked, gets most of the value available. A team with an expensive licence and no rules gets the invoice and the risk. Spend on rules and habits before you spend on software.
If two experienced people would decide differently, do not automate the decision
That is judgement, and it belongs with people. Automate the gathering, the drafting and the summarising around the decision, and leave the decision where it is. Most disappointing AI projects we get called into broke exactly this rule.
If this sounds cautious, it is caution that comes from running these systems with our own customers on the line. The businesses that do well with AI are rarely the ones that moved first. They are the ones that moved deliberately.
Frequently asked questions
We have never used AI at all. Is that a problem?
No, and you are not behind. Plenty of the businesses we work with are starting from zero, and starting from zero with good advice beats a year of bad habits. The advisory session assumes no knowledge and involves no jargon.
Is an independent AI consultancy better than a vendor?
It is when the vendor only sells one answer. We resell nothing and take no commission, which is unusual in this market and worth asking any consultant directly. We recommend whatever fits your workflow and budget, and sometimes that is a free tool used carefully.
Is our customer data safe if we use AI?
It can be, with the right habits. The main risk in a small business is staff pasting sensitive information into public tools that may keep what they are given. A short policy, a little training and sensible tool choices remove most of that risk, and we set all three up with you.
Do you write the AI policy for us?
With you, rather than for you. A policy written in your words about your actual tools gets read and followed. A template downloaded from the internet gets filed. It is usually a page long and done within a session.
Can you train our whole team?
Yes. Hands-on sessions built around your real tasks, in person across Greater Manchester or remotely UK-wide. Some businesses train everyone in the basics, others start with a small group and spread it. Both work.
What does the advisory session not cover?
It is one hour on one defined challenge, so it is not a full review of the business. If the question turns out to be bigger than an hour, we say so and recommend the assessment or an implementation plan rather than rushing an answer.
Do we have to use you for the implementation?
No, and the plan is deliberately written so you do not. It is detailed enough for your own people or a third party to build from. If you do want us to build it, we quote that separately with real numbers.
Will AI replace our staff?
In a small business, almost never. What it removes is the part of every job nobody wanted: retyping, chasing, summarising, drafting from scratch. In practice it postpones a hire rather than replacing a person, and we recommend saying exactly that to your team early.
What if the tools change six months after the advice?
They will. The principles we work to, messy input, human approval, clear rules, boring first projects, have outlived every tool cycle so far. The Ongoing AI Partner arrangement exists for exactly this, and one question by email does not need a retainer. Ask it.
Why choose a Manchester AI consultancy over a London firm?
Because proximity and evidence beat prestige. We are our own first client. Every system we recommend has run inside a real trading business with real customers, and we have paid for our own mistakes rather than yours. Ask any consultant what they run in their own company. The answer is informative.
Where this connects
Consulting is the front door. These are the rooms it usually leads to.
- AI Opportunity AssessmentThe structured version of "where should we start", with five prioritised opportunities. Founding price £99.
- Business AutomationWhen the plan says build, this is what building looks like, priced honestly.
- AI ReceptionistThe AI most small businesses should meet first: the one answering the phone.
- Content and CampaignsAI-assisted, human-approved marketing that sounds like you wrote it.
- How the platform fits togetherWhere implemented AI actually lives, and how the pieces share data.
- What AI automation can actually doThe plain-English primer, with no pricing attached.
Bring us one question
Not a strategy, not a transformation. One defined challenge, one hour, £149, and you leave with a written answer and the next three moves. If the honest advice is "do nothing yet", that is what the summary will say.
All prices plus VAT. Custom development scoped and quoted separately.