When You Shouldn't Use AI in Your Business
AI is the wrong tool more often than its promoters admit. Seven situations where you are better off with a rule, a form or a person.
We build AI systems for a living, and we talk clients out of them regularly. Knowing where AI does not belong saves money and prevents the kind of failure that puts a business off useful technology for years.
1. When a simple rule would do
If the task can be described as "when X happens, do Y", you do not need AI. Use ordinary automation. It costs less, behaves the same way every time and is easy to fix.
Sending a confirmation after a booking does not require intelligence. It requires a trigger and a template.
2. When a mistake is expensive or irreversible
AI systems make errors. Usually they are small. Occasionally they are confident and completely wrong. That is acceptable for drafting an email you will read before sending. It is not acceptable for:
- Final quotes and binding prices
- Refunds and payments
- Medical, legal or financial advice
- Safety instructions
- Anything that goes on a contract
AI can still help in these areas by preparing a draft or flagging issues. The decision and the signature should stay with a person.
3. When the customer needs a human
Some conversations are about more than exchanging information. A complaint, a bereavement, a customer who is frightened or furious. Being met with software in that moment makes things worse.
Design every customer-facing AI with a quick route to a person, and send the difficult conversations there on purpose.
4. When you cannot check its work
If nobody in your business is able to judge whether the AI's output is right, you should not rely on it. An owner who cannot read a contract should not use AI to review contracts. The tool amplifies competence; it does not replace it.
5. When your own information is a mess
An AI assistant answers from what you give it. If your price list is out of date, your policies contradict each other and half of how you work lives in one person's head, the AI will faithfully reproduce the confusion at scale.
Fix the information first. Often that alone solves the problem you wanted AI for.
6. When the task is rare
A job you do twice a year does not justify building anything. The setup effort will never be repaid, and the system will have drifted out of date by the next time you need it.
7. When it involves sensitive data you have not thought through
Health details, payment information, records about children, confidential client matters. Before any of that touches an AI tool, you need to know where the data goes, who can see it, how long it is kept and whether your obligations allow it.
That is a question for the provider's terms and, in regulated fields, your own advisor. The US National Institute of Standards and Technology publishes a useful general framework for thinking about these risks in its AI Risk Management Framework.
A quick decision table
| Situation | Better choice |
|---|---|
| Same steps every time | Rule-based automation |
| High stakes, low volume | A person, perhaps with an AI draft |
| Emotional or sensitive conversation | A person |
| Nobody can verify the output | Do not automate it |
| Messy source information | Fix the information first |
| Happens a few times a year | Do it by hand |
What is left
After ruling all that out, plenty remains: answering routine calls, replying to common inquiries, sorting messages, extracting data from documents, drafting first versions. Those are frequent, similar each time, cheap to check and easy to hand to a person when something unusual comes up.
That is where to start. Practical ways small businesses can use AI lists the best candidates.
