AI automation for small businesses starts with a useful question: which part of the work keeps getting stuck? For a business that depends on enquiries, the answer may be the information collected before anyone can give a helpful reply. A clearer form, a reliable handover or a checked summary could be a better first investment than another subscription.
The AI news is encouraging. Your starting point still needs to be specific.
On 22 October 2026, Google highlighted new Australian research into AI adoption among small and medium businesses. The examples included everyday operations such as quoting, scheduling and ordering. The Google announcement concerns Australian businesses and research supported by technology providers. Its headline estimates are not a forecast for your own UK business.
A separate OpenAI case study published on 20 October describes a US restaurant operator using AI for several administrative tasks. In the scheduling example, a manager checks and adjusts the proposed schedule before entering it into the business’s system. These are vendor-reported experiences from a particular business, rather than independently verified results we can promise you.
The useful question for a small team is what a tool would change in its own working day. Our suggested starting point is to replay three recent enquiries and look for the point where progress stopped.
Replay three enquiries before choosing a tool
Choose three typical enquiries that have already reached a conclusion. For each one, trace the journey from the first message to the reply, booking or quote. You can do the initial exercise yourself without uploading customer information anywhere.
- What did the customer tell you at the start?
- What did you have to ask before you could help?
- Who was responsible for the next reply?
- Where did someone copy the same details into another system?
- What made the enquiry ready for a decision?
Mark each pause. Was information missing, was nobody clearly responsible, or did someone need time to read a long message? Those are different problems, so they may need different solutions. Avoid choosing software until you can explain which pause you want to remove.
A fictional example: the garden enquiry that needs four more messages
Imagine a landscaping business receives: “Can you quote for my garden?” Before deciding whether to arrange a visit, the owner needs to know the location, the kind of work and the customer’s preferred timing. A photo may help, but it should not become a barrier for someone who cannot provide one.
A short form could ask for the postcode, the type of job and a brief description, with an optional photo. A customer asking about maintenance would see relevant questions; someone planning a new patio would see a different follow-up.
That change does not automatically need AI. If the missing details are predictable, you may simply need to improve your existing enquiry form. Collect what you need for the next step, while leaving the detailed specification for a conversation. This is an illustrative example, not a Panda customer result.
Does the job need a better form, automation or AI?
| Where the work gets stuck | A sensible option to investigate |
|---|---|
| The same essential details are missing | Clearer questions and conditional form fields |
| Details are repeatedly copied into a customer record | A rules-based connection between the form and your CRM |
| Long, varied messages take time to understand | An AI-assisted summary for someone to check |
| The next step needs a quote or professional judgement | A named person who reviews the information and decides |
These options can work together. A form might collect the facts, a connection might create a customer record, and AI might prepare a short summary. Each step should have a clear purpose. Our AI and workflow automation service starts by understanding the process and agreeing where people take over.
Test one improvement against the whole job
Decide what a useful result looks like before starting a trial. “The owner can decide whether to arrange a visit without chasing the postcode” is specific enough to check. “We want to use more AI” gives you much less to work with.
- Choose one type of enquiry. Keep the first trial narrow enough to understand what changed.
- Record the current effort. Include reading, chasing missing details, copying information and correcting mistakes.
- Try the smallest useful change. That could be a revised question, a notification rule or a draft summary.
- Check awkward cases. What happens when an answer is missing, a customer changes their mind or a connection fails?
- Compare complete enquiries. Count the work needed to reach a useful next step, including checking and rework.
A faster draft is only part of the picture. If someone spends longer correcting it, the job may not have improved. Include subscriptions, setup and ongoing supervision when deciding whether a change earns its place.
Do you need a new website to do this?
Sometimes the existing form or business software can be improved. Sometimes an integration needs custom work. A rebuild should follow from the requirements and the condition of the current site.
If the brief involves complex form journeys, customer accounts or several connected systems, bespoke website functionality may be worth discussing. We agree the scope, costs and responsibilities before work begins; you can see how we plan and build a project.
Keep ownership of the reply
Agree who checks uncertain summaries, who handles failed connections and who can pause the process. Start tests with made-up or appropriately prepared examples, and understand what information a proposed tool receives before connecting it to real enquiries. Advice, prices and commitments need the review arrangements you have agreed.
For your next step, take those three enquiries and write down the one delay that keeps recurring. That gives you a concrete brief for improving the process. Tell us where your enquiries get stuck, and we can help work out which changes are worth exploring.