Hands-on AI training and ongoing process improvement, with before-and-after numbers.
I work with your team on the tasks and processes that matter most. Using ChatGPT or Claude, I help them build automations around their own work and measure the difference they make.
★★★★★
“I didn’t just leave with something built for me. I understood how it worked and felt confident improving it myself afterwards.”
Liz AshtonFounder, Symbioss
★★★★★
“We cut the time spent finding and qualifying prospects from eight hours a week to two, while keeping final control over every message.”
Will ThompsonFounder, Goodwill
★★★★★
“I learned how to use AI on the work I actually do. Now it handles the research and admin, without taking away the judgement that makes the business mine.”
Harriet MurphyFounder, HOTEL, MIKE
The problem
Everyone’s using AI. Few get real time back from it.
How most people use it
Useful, but it doesn’t yet feel like actual automation.
- Every task starts with a blank chat, or pasting information in.
- Answers go back and forth until the tone, format or brand is right.
- Next week, it all starts again in a new chat.
What people rarely use
Where the real power and time saving are.
All built into the tools you already pay for.
Services
For each person’s own tasks
AI Training
£950 plus VAT
Two half-days · up to six people
Move from chat to reusable automations: a second brain with your context, skills for the jobs you repeat, and scheduled tasks that run themselves.
Each attendee saves half a day a week, on average.
More on AI Training →For the work your team shares
AI Programme
£3,000 a month, plus VAT
Ongoing · run every month
One backlog of improvements to the processes your team shares: mapped, built and measured before and after.
Every change measured: time, cost, quality and the outcome you care about.
More on AI Programme →Not sure where to start? Book a free 30-minute call and I’ll help you choose.
Everything I build sits in the tools you already use.
I teach
ChatGPT
ClaudeConnects to 2,000+ more tools through ChatGPT and Claude’s own connectors
Works with
AI Training
Fazed Media cut race-plan time from five hours to under one.
Sarah Fasey runs social media for elite athletes. In one hands-on day she built a second brain, connected her tools and turned recurring work into reusable automations.
14 hours saved a month
Read the full storyAI Programme
Hotel Mike grew from two new retailers a month to six.
We redesigned the whole wholesale process, from finding stores to orders and invoicing. Harriet kept final control over every decision.
3× new retailers a month
Read the full story
Michael Atkins
CairnWorks is just Michael. A solo consultancy, for now.
Certified by AnthropicTeaching AI Fluency. Verify (opens in a new tab)Certified by OpenAIAgents and Workflows. Verify (opens in a new tab)Four ways to improve how work gets done with AI.
| Do it yourself | Hire a full-time AI expert | Have AI built for you | cairnworks | |
|---|---|---|---|---|
| Cost | Licences only, but progress costs internal time. | Salary, recruitment and onboarding. | A project fee, with further costs when requirements change. | A fixed price for Training. A monthly fee for the Programme. |
| Speed to value | Quick to start. Slower to find where AI makes the biggest difference. | Usually months to recruit, then time learning your business. | Time is needed to scope and build before anything goes live. | Working automations across two half-days. Then a shared process improved every month. |
| Measurement | Often not measured unless someone deliberately tracks it. | Depends on the person and how success is managed. | Usually shown at handover. Longer-term impact may not be tracked. | Measured against your own work and where you started. |
| What happens after | Depends on someone finding the time to continue. | Capability stays inside, but may sit with one person. | Future changes usually depend on the supplier. | Your team understands what was built and can keep improving it. |
Do it yourself
Cost
Licences only, but progress costs internal time.
Speed to value
Quick to start. Slower to find where AI makes the biggest difference.
Measurement
Often not measured unless someone deliberately tracks it.
What happens after
Depends on someone finding the time to continue.
Hire a full-time AI expert
Cost
Salary, recruitment and onboarding.
Speed to value
Usually months to recruit, then time learning your business.
Measurement
Depends on the person and how success is managed.
What happens after
Capability stays inside, but may sit with one person.
Have AI built for you
Cost
A project fee, with further costs when requirements change.
Speed to value
Time is needed to scope and build before anything goes live.
Measurement
Usually shown at handover. Longer-term impact may not be tracked.
What happens after
Future changes usually depend on the supplier.
Cost
A fixed price for Training. A monthly fee for the Programme.
Speed to value
Working automations across two half-days. Then a shared process improved every month.
Measurement
Measured against your own work and where you started.
What happens after
Your team understands what was built and can keep improving it.
Questions, answered.
Do we need technical experience or new software?
No. If your team can use ordinary business software, they can take part. I handle the setup and explain anything unfamiliar as we go. Training mainly uses the Claude or ChatGPT desktop app, which comes with a normal subscription.
Which is right for us: AI Training or the AI Programme?
Most teams start with AI Training, so everyone can automate their own work. The AI Programme is for teams already doing that, who want the processes they share to improve too. A free 30-minute call helps work out which fits.
How is the work tailored to our business?
Before Training, everyone completes a short interview about their role, recurring work and tools. The Programme starts by mapping how your shared processes really run today.
Is it safe to use our client information?
It depends on your plan, so that’s where we start. On the business plans (ChatGPT Business or Enterprise, Claude Team or Enterprise), your content isn’t used to train their models by default, and both offer a data processing agreement that covers UK GDPR. On personal plans, whether chats are used to improve their models is mostly down to a setting each person controls, and it isn’t always off. I check this with you before anyone connects anything.
Where data is stored depends on the provider and plan: mostly the US, with UK or EU options on some enterprise plans. If your client contracts say where data must be held, we check that first too.
Connecting your email, files or other apps has its own settings and terms, on both the AI side and the app side, so we check those too. Everything is set up in your own accounts.
The one thing I hold is what your team puts into my training app: their interview answers before the day, and the task maps they fill in during it (the steps, tools and time each task takes). It asks how you work, never for client information. It’s kept in the UK and EU, under a data processing agreement, and deleted 12 months after the training.
How do you measure whether it was worthwhile?
We agree what matters first: time, cost, quality or the outcome you care about. I record the baseline and measure the change against it.
What happens after the engagement ends?
After AI Training, you keep everything built and the instructions to carry on without me, plus three weeks of email support for questions or problems with what we built.
When the AI Programme ends, you keep everything built, the backlog and the measurements, so your team can keep improving the work on its own.
Find out what good looks like, and the steps that would get you closer.
The check gives you next steps straight away. On the call, I show you what teams getting real value do differently, and give you three things to change.