AI consulting for SMEs.

Every software vendor now says it has AI, and most business owners have been told they should be doing something with it. This is advice before any building: we look at how your business actually runs and tell you where AI would save real time or money, where ordinary automation would do the job better, and where nothing needs to change.

If the answer is to build something, our AI automation page covers that work. This page is about deciding first.

• Reading messy documents. Supplier invoices, delivery orders and forms that look different every time.
• Sorting incoming messages. Working out what an email or chat message is asking for and sending it to the right person or queue.
• Drafting, with a person approving. Replies, quotes, product descriptions and summaries written in your tone, sent only when someone says so.
• Finding things in your own documents. Answering staff questions from your manuals, price lists and past jobs.
• Pulling structure out of free text. Turning notes and messages into fields a system can use.

• Rules you can write down. If the logic is “when X, do Y”, plain code is cheaper, faster and predictable.
• Decisions where a mistake is costly. Payments, medical, legal or safety decisions need a person, with AI at most preparing the case.
• A chatbot for its own sake. Most customers want a fast answer or a person, and a poor bot costs goodwill.
• Bad or missing data. AI does not fix records nobody keeps; the first project may be getting the data in order.
• Low volume. A task done a few times a month rarely justifies the setup and the running cost.

1. Pick one task

   One repetitive job with a clear before-and-after, such as reading supplier invoices or sorting customer emails.

2. Test it on your real examples

   We run a sample of your own documents or messages through it, with personal data handled carefully, and measure how often it is right.

3. Look at the misses

   What it got wrong, why, and whether a person checking the unsure cases makes it safe enough.

4. Decide with evidence

   You see the results and the expected running cost, then choose to build it properly, change course or stop. Stopping is a fine outcome.

Using AI means sending some of your data to a provider’s model. We pick providers and settings that do not train on your data, keep storage in accounts you control, send only what a task needs, and keep a person in the loop wherever a mistake would matter. We plan with the Personal Data Protection Act in mind and tell you plainly where the risks are.

For many SMEs the bigger risk is staff pasting customer data into free AI tools. Part of a consultation can be a simple, sensible policy for that.

• A short written view of where AI fits in your business and where it doesn’t, in plain language.
• For the most promising idea, a pilot plan or pilot results, with the running cost estimated.
• An honest recommendation, which may be ordinary automation, an existing tool you already pay for, or doing nothing yet.

• Automation & AI Building the automation once you have decided.
• IT consulting Advice on systems and vendors in general.
• EDGE grant explainer What the grant says about digital projects.
• FAQ Cost, timelines, who owns the code.