01
AI readiness review
We look at your tasks, tools and data, and at how your team already uses AI, then write up where Claude could help, what to tackle first and what to leave alone.
AI with Claude · Southampton & the UK
Practical AI consulting with Claude, for small businesses.
We're an AI consultancy based in Southampton, and we predominantly use Claude, made by Anthropic. We help small teams work out where AI genuinely saves time, set Claude up safely, train people to use it well, and build Claude-powered tools into the systems they already use.
What we help with
From first questions to everyday use.
Most teams have tried an AI chat tool and aren't sure what to do next. As a Claude AI consultant for UK small businesses, we help you get past that point with a plan that fits how you work.
02
Choosing where AI helps
AI is good at drafting, summarising and sorting. It's poor at anything needing guaranteed accuracy or accountability. We help you tell the difference.
03
Setting up Claude for your team
Getting your team working with Claude: shared projects that hold your own reference documents, set up and looked after by us, with no licences for you to manage.
04
Training and a safe-use policy
Hands-on sessions built around your team's real tasks, plus a short written policy: what can go into Claude, what can't, and what must be checked.
05
Prompt libraries and workflows
Tested, reusable prompts for the jobs you repeat each week, such as reply drafts, meeting notes and report summaries, so results are consistent whoever runs them.
06
Connecting your documents and tools
Giving Claude safe access to the documents and systems it needs for a task, using its built-in connections where they exist and custom ones where they don't.
Built in-house
Custom Claude-powered tools for your business.
Some jobs are better served by a small, focused tool than a chat window. Claude has an API, so our team can build it into software that does one thing well, with your rules, your documents and a person approving the result.
Our team designs and builds these tools in-house, led by Marcel, who is your primary contact for software and AI; if he's unavailable at short notice, another member of our team is on hand. They can live inside your website, your HR platform or a private web page for your staff.
- Enquiry triage. Reads each new enquiry, tags it by type and urgency, and drafts a reply for someone to approve.
- Quote and proposal drafting. Builds a first draft from the client's brief, your price list and past work. See business proposals.
- Internal knowledge assistant. Answers staff questions from your own procedures and documents, and shows where each answer came from.
- HR policy assistant. Answers everyday questions about holiday, absence and expenses from your handbook, and passes anything sensitive to a person.
- Inside your website. Smarter enquiry forms or customer-facing tools, built by our web design team.
Honest limits
What AI can't do, and how we plan for it.
AI tools, Claude included, make mistakes. They can misread a document, get a number wrong or state something false with total confidence. That doesn't make them useless, but it does decide how they should be used.
We design every setup and every tool on the assumption that the AI will sometimes be wrong, so a person stays in the loop wherever it matters.
- A person reviews the output. Drafts are drafts until someone who knows the subject has checked them.
- Not for regulated decisions. We don't use AI to make legal, financial, medical, credit or employment decisions, or to give regulated advice.
- No decisions about individuals. AI can help organise HR information, but people decide about people.
- Sources you can check. Where possible, tools show which document an answer came from.
- No overpromising. If a task isn't a good fit for AI, we'll tell you, even if that means less work for us.
How it works
How an AI project runs.
Every meeting is a video call, booked at a time that suits your team, so we work the same way with businesses in Hampshire, Bournemouth, the Isle of Wight and anywhere else in the UK.
-
Readiness chat
We learn how your team works, which tasks eat the most time and what you've already tried.
-
Plan and fixed quote
A short written plan with priorities and a fixed quote for our work, with Claude included. Nothing is due upfront.
-
Set up, train and build
We set up Claude, run training, write the safe-use policy and build any custom tools, testing them on your real tasks.
-
Review and refine
After a few weeks of use we check what's working, tune prompts and tools, and drop anything that isn't earning its keep.
Data and safety
Your information stays under control.
Using AI safely is mostly about deciding what goes into it. We set that out clearly, in writing, for your team.
- Business terms. We use Claude on Anthropic's commercial terms, under which your data isn't used to train its models by default.
- Only what a task needs. Tools we build send Claude only the information required for that job.
- Clear rules for staff. Your safe-use policy says plainly which kinds of information should never be pasted in.
- Transparent. Our privacy policy explains how we handle personal data.
How we price
Clear costs for our work and for Claude.
Every engagement is quoted individually, from a single readiness review to a custom-built tool. You get a fixed written quote before we start, and nothing is due upfront. Changes to tools we've built, after launch, are charged at our published hourly rates: £50 an hour for jobs of an hour or less, £30 an hour beyond that.
Claude itself: included. We provide Claude as part of the work, so there are no licences, subscriptions or usage bills for you to manage.
Questions
Questions about AI with Claude.
Why Claude rather than ChatGPT or Copilot?
We predominantly use Claude because it's the tool we trust most for reading long documents, working with data and writing clearly, and because Anthropic puts a strong emphasis on safety. Plenty of good tools exist, though. If your business is already committed to another one, we'll talk about whether switching is worth it rather than assume it is.
Are you an Anthropic partner?
No. We're an independent business that uses Claude in our own work and builds on Anthropic's API. We don't earn anything from Anthropic for recommending it.
Do we need to buy Claude licences?
No. We provide Claude as part of our work — for data insights, for the tools we build and for the workflows we set up for your team — so there's nothing for you to buy, license or keep track of. It's all covered in your quote.
Will AI replace people on our team?
That isn't how we approach it. The aim is to take the repetitive parts out of people's work, like first drafts, summaries and sorting, so they spend more time on the parts that need judgement and a human touch.
Can Claude work with the tools we already use?
Often, yes. Claude can connect to a number of common tools, and where there's no ready-made connection, our team can build one using Claude's API. We'll check your systems during the readiness review.
Can AI help us understand our business data?
Yes, with a person checking the results. See data analysis and reporting for how we use Claude to turn your figures into plain-English reports.
Ready to put Claude to work?
Tell us which tasks take up your team's time. We reply within one business day.