Where small businesses lose time
Most owners lose hours every week to the same few jobs: replying to enquiries, chasing reviews, building reports, drafting posts and copying data between tools. These jobs are repetitive, rule-based and easy to forget, which is exactly where automation pays off.
How I choose what to automate
I start with a workflow audit. We list the recurring tasks, how long each takes, and what a mistake would cost. Then I pick one or two that are frequent and low-risk and build those first.
My rule is that every automation needs a measurable job. If I can't say which number it moves, such as response time, review count or hours saved, it is a toy and I don't build it.
Examples of what I build
- New enquiry arrives, gets logged in a sheet or CRM, receives an instant acknowledgement, and reminds you if nobody has replied within a day.
- After each sale, a polite review request goes out by WhatsApp, SMS or email with your Google review link, feeding your Google Maps ranking.
- A weekly report that pulls numbers from Search Console, Analytics and your Business Profile into one summary.
- A content pipeline: brief, AI first draft, human edit, then scheduling (see content marketing).
- Research agents for jobs like competitor scrapes, with the output reviewed before anyone acts on it.
Why a human stays in the loop
AI can be confidently wrong, so every workflow I build has a review point before anything reaches a customer. Sensitive messages arrive as drafts for you to approve. If your data must not be processed by third-party AI tools, tell me at the start and I will scope the work around it. You can read how I handle information in my privacy policy and terms of service.
The tools
I pick one tool for each step rather than forcing one tool to do everything. Claude handles long-context work such as strategy notes and SOPs, ChatGPT is quick for drafts and small data jobs, Manus runs autonomous agent tasks, and Notion and Slack carry the hand-offs. Follow-ups feed straight into lead generation, so the extra enquiries are tracked.
What a first project looks like
A first project is deliberately small, so you see the value before committing to more:
- A call to list your repetitive tasks and pick the one worth automating first.
- A build using your real tools, tested with real examples before anything goes live.
- A week of watching it run, with fixes where it needs them.
- A short guide to what it does, so anyone on your team can understand it.
From there we can add the next workflow, or leave it as it is. There is no pressure to automate everything.
