AI Audit
Most company owners can feel the AI opportunity now. Beyond the rhetoric, what's actually possible with today's tools in a business like yours, this quarter?
AI Audit
That's what an AI Audit answers (you might also know it as an AI Readiness Assessment). We look at how your people actually work day to day, and where their hours are going on tasks AI could now handle. Then we look outside, because your customers are starting to arrive via agents too. How agent-ready is your website? When an LLM lands on it, does it understand how to interact with what you sell? We map both sides, write it up, and hand back a prioritised plan you can act on without us.
The shape of the audit, in detail.
Inside the team
How your people work today, and where AI fits in.
- Where do the hours go?
- We sit with the people doing the work, watch a typical work-pattern, and identify what they don't want to be doing. The repetitive shaping, the spreadsheet reformatting, the knowledge lookups that should take seconds.
- What's already in play?
- Claude, ChatGPT, Copilot, Gemini. What your team actually uses (or wishes they could), what your security policy says they should, and where the disconnect is.
- What data sits where?
- Drive, Slack, your CRM, Shopify, GA, GSC, the analytics warehouse. Where AI could plausibly read or write, and where it shouldn't.
- Who could own this?
- We look for the AI owner inside the team. Sometimes that's a person with capacity in an existing role, sometimes it's a role you'll need to define before any kit lands.
Outside the business
Your customer-facing surface, through the eyes of an agent.
- How agent-ready is your website?
- When an LLM crawler arrives, can it read what you sell? Is the structured data rich enough for an agent to act on it without guessing?
- What does AI Search say about you?
- We check what ChatGPT, Claude, Perplexity and Google's AI Overviews actually return when someone asks about your business, and where the gaps are.
- Are AI crawlers blocked at the edge?
- Plenty of sites still have ClaudeBot or GPTBot rejected at the firewall by accident. We check, and recommend a position (allow, block, or per-bot).
- Could WebMCP help?
- For service businesses (booking, search, configure, quote, apply) exposing the action layer to agents directly is a real lever. We assess whether it's worth doing for you now or later.
The output
What you walk away with.
- A written audit, not a slide deck
- Prioritised by impact and effort, so the conversation in the boardroom is grown-up.
- A recommended next phase
- The Build, an AI Pilot, AI Training, or "you're not ready yet, here's how to get ready". We're honest about which.
- Something your team can act on without us
- If the audit is where the engagement ends, that's fine. Plenty of teams come back a quarter later when the timing is right.
The teams that get the most out of it.
Marketing, e-commerce, data and ops teams in SMEs. Especially the ones where the question on the table is "what should we do about AI?" and the honest answer is "we're not sure yet". This engagement turns that into a written plan you can act on, take to the board, or sit on for a quarter while you finish something else.
The shape of the engagement.
A short, focused engagement with the people closest to the work. We meet your team, watch how a typical week goes, ask the kind of questions an outsider can ask without it being awkward, and write it all up. The output is a prioritised plan, not a deck. Something your team can refer back to a year later and still recognise.
What the next conversation looks like.
If the audit points to a clear build, the next conversation is The Build (which itself begins with a sandboxed prototype, so a smaller-scope feasibility test is built into the engagement). If it points to "we need to upskill before we touch this", that's AI Training. The audit is honest about which is right for you, and a fair number of audits end with "you're not ready yet, here's how to get ready".