If you run a small or medium-sized business, it is easy to feel that “we need to do something with AI” has become another item on the management to-do list.
The awkward bit is knowing what that “something” should be.
Some businesses need training. Some need consultancy. Some are ready for a practical AI system or workflow build. Many need a mixture, but not necessarily all at once.
Getting this right matters because the wrong first step wastes time and money. A half-day training session will not fix a broken admin process. A custom AI system will not help much if nobody in the business understands where it fits. And a broad strategy document is not enough if the team really needs a working tool that saves time every week.
Here is a simple way to think about the difference.
AI training: when the main problem is confidence and basic capability
AI training is useful when your team needs to understand what AI can and cannot do, how to use it safely, and where it might fit into everyday work.
Typical training topics include:
- Writing better prompts
- Checking AI output rather than trusting it blindly
- Using AI for first drafts, summaries, research and admin
- Handling confidential information sensibly
- Spotting tasks where AI is a poor fit
- Setting simple internal rules for staff use
Training is usually the right starting point if people are curious but hesitant. Perhaps a few staff have tried ChatGPT, Copilot or Gemini, but usage is patchy. Some people are enthusiastic. Others are worried about quality, privacy or looking foolish. In that situation, training can create a shared baseline.
The limitation is that training is not implementation. It can make people more capable, but it does not redesign a workflow for you. If the business problem is “we spend too long triaging enquiries” or “our reporting process is slow and inconsistent”, training may help the team understand the opportunity, but it will not automatically create the new process.
Use training when the question is: “How do we help our team use AI sensibly?”
AI consultancy: when the main problem is deciding what to do
AI consultancy is useful when you can see potential, but you are not sure which opportunities are worth pursuing.
This is often the most sensible step for SMEs because the challenge is rarely a lack of AI tools. The challenge is choosing the right problem.
A good consultancy conversation should look at your business, not just the technology. Where is time being lost? Which tasks are repetitive? Which decisions depend on lots of information? Where do mistakes create risk? Which processes already have clear rules? Which systems hold the data? Who needs to remain in control?
The outcome should be practical. Not “AI transformation” in vague terms, but a shortlist of opportunities such as:
- An assistant that helps draft and classify customer emails
- A workflow that turns meeting notes into actions and follow-ups
- A research process that gathers public information and produces a human-checkable summary
- A local/private AI setup for sensitive internal documents
- A small agent that monitors sources and flags relevant changes
- A better internal process for using AI without leaking data or creating quality problems
Consultancy is especially useful when there are several possible directions, and you need to compare value, risk, cost and complexity before building anything.
In my own Cheltenham Times experiment, the valuable lesson was not simply that an AI agent could write articles. It was that the agent needed rules, correction, source checks, cost control and human judgement. The more useful parts came when the work moved beyond generic content and into practical local assets, such as pages that could be refreshed and improved over time. That is the sort of distinction consultancy should help you make: not “can AI do this?” but “is this the right thing for AI to do, and how should we manage it?”
Use consultancy when the question is: “Where could AI create real value in our business, and what should we do first?”
AI systems and workflow builds: when the main problem needs a working process
An AI system or workflow build is the right route when you have a clear use case and need something operational.
This might be a lightweight internal assistant, a semi-automated admin process, an AI-supported CRM workflow, a reporting tool, or an agent that checks information sources and prepares drafts for review.
The key word is “system”. This is not just opening a chatbot and asking it questions. A useful AI system usually involves several parts working together:
- A defined trigger, such as a new email, file, task or scheduled check
- Access to the right information
- Clear instructions and guardrails
- A structured output
- A human review step where needed
- Logging, measurement and cost control
- A way to improve the workflow as you learn
For example, my AI-first CRM was not just a chatbot added to a contact database. The idea was to explore what happens when AI agents are built into the working day, helping with tasks, splitting larger work into smaller steps, supporting different roles, and connecting to local machines where appropriate. That kind of work needs design and testing, not just enthusiasm.
For most SMEs, the first AI system should be modest. Start with one workflow where the inputs are clear, the output is easy to check, and the time saving is visible. You do not need to automate half the business. You need one useful improvement that proves the approach.
Use a system build when the question is: “Can we turn this specific process into a reliable AI-supported workflow?”
How to choose the right first step
A simple rule is this:
If your team does not understand AI well enough to use it safely, start with training.
If you understand the basics but do not know which opportunities are worth pursuing, start with consultancy.
If you already have a clear, valuable, repeatable process in mind, consider a workflow or system build.
There is no prize for jumping straight to the most complex option. In fact, the safest AI projects are often the ones that start small, prove value, and then expand.
A Cheltenham accountancy firm, marketing agency, estate agent, manufacturer or professional services firm might all use AI differently. The right answer depends on the work, the risks, the data, the team and the commercial objective.
The useful question is not “Should we use AI?”
It is: “What is the first business problem where AI could help enough to be worth the effort?”
A practical next step
If you are unsure whether you need AI training, AI consultancy, or a small AI system, start with a conversation about the work itself.
What takes too long? What gets repeated? What do customers ask for again and again? Which internal processes rely on copying, checking, rewriting or chasing information?
Those answers usually point to the right starting point.
If you would like help identifying the first sensible AI opportunity in your Gloucestershire or Cheltenham business, book a short AI consultancy call with Stuart Cole Consulting. We can look at your current workflows, separate useful ideas from hype, and work out whether training, consultancy or a practical system build is the right next move.
