One of the most useful prompts is: “Before answering, ask me what you need to know.”
The Polite Problem With AI
AI is often too willing to answer. It will usually produce something even when the brief is thin, the facts are incomplete or the objective is unclear. That can be helpful for brainstorming. It is less helpful when the output will influence a customer, a process, a proposal or a decision.
In business, an answer based on hidden assumptions is not merely imperfect. It can be actively misleading.
Make the Gaps Visible
A simple technique changes the conversation: “Before you answer, ask me up to five questions that would help you produce a better result.” This forces the tool to surface missing information. It also helps the user notice details they may have assumed were obvious.
The questions may concern audience, purpose, constraints, tone, evidence, examples, deadlines, decision authority or what the business has already promised.
Why This Works for Small Businesses
SMEs often run on tacit knowledge. The owner knows which customers are sensitive, which suppliers are awkward, which processes contain exceptions and which words should not be used in a particular context. That knowledge is obvious to the owner precisely because they carry it every day. It is not obvious to an AI tool.
Asking for questions first is a way of moving knowledge out of the owner’s head and into the brief.
When to Use the Technique
Use it whenever the task is more than routine wording. It is especially useful for proposals, complaint responses, policies, job adverts, customer communications, process design, website copy and planning documents.
For example, before drafting a customer complaint response, AI might ask: Is the customer long-standing? Has the business accepted fault? Is compensation available? What outcome do you want? Are there phrases or promises to avoid? Those are precisely the questions a sensible colleague would ask — the same instinct behind asking in steps.
Ask for Risks as Well as Questions
A stronger version of the technique is: “Before answering, ask me any essential questions. Also list the assumptions you would otherwise have to make.” This is valuable because some gaps may not need a question, but they do need acknowledgement.
For example, AI might say it is assuming the customer is based in the UK, that the product is still under warranty, that the business has authority to offer a replacement, or that the reader is already familiar with the service. Once the assumptions are visible, the business can accept, correct or remove them.
Do Not Let Clarification Become Theatre
There is one caveat. AI can ask too many questions, or questions that do not matter. The point is not to turn every email into a consultation. Set limits. Ask for the five most important questions, or only the questions that would materially change the answer.
For quick, low-risk tasks, one or two clarifying questions may be enough. For important work, a fuller clarification stage is sensible.
Practical Steps You Can Take Today
- Add a standard line to important prompts: “Ask me any questions you need before drafting.”
- For proposals and customer replies, ask AI to identify missing information that could change the tone or recommendation.
- Ask it to separate essential questions from nice-to-have questions.
- Use the questions as a training tool. They show staff what a good brief should include.
- Save the best clarification questions and turn them into a pre-AI checklist.
Questions Are a Sign of Control
Some people feel that asking for clarification slows the process down. In reality, it prevents avoidable rework. A rushed answer can be more expensive than a careful question.
AI will not naturally understand your business, your customers or your risks. But it can be prompted to ask before it assumes. That single habit can improve the quality of almost every serious AI-assisted task.
