There's a version of this article that lists ChatGPT, tells you to use it for email writing, and calls it a day. This isn't that article.
What follows is what I'm actually seeing work for real businesses — the ones that have moved past the hype and are getting genuine, measurable value from AI.
1. Automating customer enquiry triage
The most common time-sink I find when I audit businesses is the inbox. Specifically, the process of reading a customer enquiry, working out what it's about, and routing it to the right person.
A well-implemented AI triage system can handle this automatically. Not just keyword matching — proper intent detection that understands what the customer actually needs and routes accordingly. Businesses doing this are saving 3–6 hours per week across their team, and customer response times have dropped dramatically.
The implementation isn't complex. It doesn't require a data science team. It requires someone who knows how to connect your email or CRM to an AI model properly.
2. First-draft document generation
Proposals, reports, case studies, job descriptions — any document that follows a consistent structure and draws on information you already have is a candidate for AI assistance.
The key word is assistance. AI doesn't replace the human who knows what the document needs to achieve. It removes the blank page problem and handles the structuring work, leaving the human to focus on what actually matters: the specific insight or decision that this document exists to communicate.
Businesses doing this well are producing documents that used to take 4 hours in under 45 minutes — and the quality is higher because the writer has more time to focus on substance.
3. Meeting summaries and action extraction
I've yet to meet a business that doesn't have a meeting problem. Too many meetings, inconsistent note-taking, action items that get lost.
AI transcription and summarisation tools have made serious progress. The current generation can produce accurate summaries with extracted action items, assigned owners, and follow-up dates — from a recorded meeting, in minutes.
The businesses doing this well have completely changed how meetings work. Fewer people need to attend because the summary is reliable. Follow-ups actually happen because the system extracts them automatically.
4. Data analysis and reporting
Most small businesses have more data than they realise, and less ability to use it than they need. Monthly reporting that used to require an afternoon with spreadsheets can now be automated entirely.
More interestingly, AI can answer questions about your data that you wouldn't have thought to ask. Pattern detection, anomaly identification, trend forecasting — at a level that used to require a data analyst.
This is one area where I'd encourage businesses to think carefully about what they actually need. A £200/month analytics tool with an AI layer often delivers more value than a £50,000 custom data science project.
5. Customer service knowledge bases
For businesses with complex products or services, a properly implemented AI knowledge base transforms how customer queries are handled. Not a simple FAQ chatbot — a system that understands natural language questions and provides accurate, useful answers from your actual documentation.
Done properly, this handles 40–60% of routine customer queries without human involvement, and the ones it escalates come with context that makes them faster to resolve.
The pattern across all five of these is the same: AI works best when it's solving a specific, well-defined problem — not when it's deployed vaguely to "use AI more."
If you're not sure where the right opportunity is in your business, that's exactly what the diagnostic tool on my homepage is designed to help with.
Or just book a call. I'll tell you honestly what I think in 30 minutes.