FREE · 3 MINUTES · NO SALES CALL
The AI Readiness Score
Eight point-and-click questions about your data, systems and processes. You get a readiness score out of 100, the gaps most likely to sink an AI project, and the first project we'd actually recommend — not the flashiest one.
Want to know what AI could save you instead? Take the 2-minute Opportunity Scan.
Where does your business information actually live?
Customer details, job history, prices, supplier terms.
How to start AI automation: the checklist
Whether you score 20 or 90, the way in is the same. Seven steps, in order — skip one and it's usually the reason a first project stalls.
- 1
Pick one process, not ten
Choose the single process that costs the most hours each week — quoting, invoicing, enquiry handling, scheduling. Automating everything at once is how projects die.
Choose your process in the workshop - 2
Count the real volume
How many times a week does it happen, and how many minutes each time? If the answer is under five hours a week, pick a different process — the maths won't pay back.
Set your weekly volume in the workshop - 3
Get the data out of inboxes
AI can't read what's in someone's head or personal inbox. The process inputs (enquiries, orders, documents) need to land in one shared system first.
Map where your inputs arrive today - 4
Connect the systems it touches
If the process spans email, a spreadsheet and your accounts package, connect those before automating. Re-keying between them is usually half the waste.
List the systems it touches - 5
Write down the steps as they actually happen
Not the official version — the real one, including the workarounds. A one-page honest map beats a fifty-page idealised one.
Walk through the real steps - 6
Agree what a human must always approve
Sending money, promising dates, quoting prices, anything contractual. Automate the preparation; keep the approval. This one decision removes most of the risk.
Mark what a human must approve - 7
Pilot for 30 days with real work
Run the automation alongside the old way on live work. Measure hours saved and errors caught. Then decide whether to roll out, fix, or stop — with evidence, not enthusiasm.
Get a phased 30-day pilot plan
Every step above is covered inside the Process Automation Workshop — walk one process through it and you get the whole checklist answered as a tailored plan.
Common questions
How do I start with AI automation?
Start with one painful, repetitive process — not with a platform or a tool. Write down how that process works today, step by step. Then automate the boring middle of it (reading, re-typing, chasing, filing) while a human keeps approving the decisions that matter. The eight questions above tell you whether your data, systems and team are ready for that first project, or whether the first job is fixing the foundations.
Is there a checklist for getting started with AI automation?
Yes — the checklist below the score covers it: pick one process, count its weekly volume, get the data out of inboxes, connect the systems it touches, document the steps, agree what a human must always approve, then pilot for 30 days. Most failed AI projects skipped one of those steps, usually the data or the documentation.
What does the AI Readiness Score measure?
How ready your business is to get value from AI and automation: the quality of your data, how connected your systems are, how well-documented your processes are, and whether your team would actually use new tools. High opportunity with low readiness means the first project should fix the foundations.
What should I automate first?
The task with the highest weekly volume and the lowest judgement required: enquiry follow-up, document intake, quote preparation, chasing paperwork, or weekly reporting. Avoid automating anything where mistakes are expensive or irreversible until you have a human-approval step that works.
Do I need to replace my existing software to use AI?
Almost never. Most good first projects connect what you already pay for — accounts, email, CRM, job management — so data flows between them, then add AI on top of that flow. Replacing systems mid-automation is the most common way projects stall.
How much does a first AI automation project cost in the UK?
A foundation piece (connecting systems, centralising data) typically runs £4,000–£12,000. A single well-scoped automation, end to end, £6,000–£15,000. A bespoke AI agent for a major bottleneck, £15,000–£35,000. Anything quoting six figures for a first project should be challenged hard.
How is this different from the AI Opportunity Scan?
The Opportunity Scan measures how much AI could save you. The Readiness Score measures how prepared you are to start. Most businesses score high on opportunity and middling on readiness — the gap tells us whether to start with a quick automation win or a data/systems foundation first.
What if I score low?
A low score is normal and fixable — it usually means your data lives in inboxes and spreadsheets rather than systems. The recommended first project will be a foundation piece (connecting systems, cleaning up one core process) that pays for itself and makes everything after it easier.
How accurate is the recommended first project?
It's a strong starting hypothesis based on the patterns we see across UK SMEs, not a specification. The £499 AI Ops Audit turns it into a fixed, written plan with exact build costs — credited against any build that follows.
Related tools: 2-Minute AI Opportunity Scan · ROI Calculator · AI Opportunity Finder · AI Automation Roadmap