The right question starts with a decision
“Monitor artificial intelligence”, “follow the competition” or “find out what is happening with batteries” are topics, not sufficiently specific intelligence assignments. They can be starting points, but they allow you to gather endless information without knowing when it is useful.
A key intelligence question (Key Intelligence Question, KIQ) steers observation and analysis towards a decision need. A solid formulation recognises at least the decision maker, the alternatives, the object, the time horizon and the conditions that would change the choice.
The question must not presuppose an answer. “Why are our competitors going to overtake us?” already contains a conclusion; “which verifiable signals indicate changes in their production capacity and what implications would they have?” invites investigation.
A five-step method
Step 1 · Define the choice. Which three plausible actions could the organisation take? Avoid “being informed”.
Step 2 · Name the decision maker. Who can make the choice and how long do they have?
Step 3 · Determine the relevant uncertainty. Which external fact do we need to understand because it would shift the balance between alternatives?
Step 4 · Formulate a neutral KIQ. Use “what”, “how”, “under what conditions” and “with what signals” before “why will it happen”.
Step 5 · Set review indicators. Which discovery would force a re-examination of the decision, even if it contradicts our initial idea?
Interactive · Builder
Live KIQ builder
Fill in the five steps. The sentence is built as you type, and below you will see whether it meets the criteria of a good KIQ.
KIQ builder: decision-maker, alternatives, horizon, uncertainty and indicators.
- A decision-maker is identified.
- There are real alternatives, not “being informed”.
- It has a time horizon.
- The uncertainty does not presuppose the answer.
- There are two observable indicators.
Company A case · simulation
Fictional caseExample · Company A · Three routes, not a single narrative
The simulated decision is between A. limited trial, B. monitoring without immediate investment, C. collaboration with specialists. The KIQ could be:
“Which evidence of technical feasibility, industrial adoption and regulatory evolution of thermal storage would justify Company A moving from observation to a trial or collaboration over the next 12 months?”
Verifiable sub-questions:
- Which results come from independent tests and under what technical conditions?
- Which industrial users show commitments beyond initial interest?
- Which regulatory changes are formally approved and which are still under consultation?
What changes the decision? A satisfactory independent pilot may favour A; evidence of collaboration with shared risks may favour C; relevant technical uncertainty and a low cost of waiting could support B. It is not a universal rule: it depends on Company A’s capacity, resources and risk appetite.
Educational simulation: fictional information and organisations.
Do it yourself
Do it yourself · Manual KIQ builder (15 minutes)
Complete this statement without using confidential data:
“Before [date/horizon], [decision-maker role] must choose between [alternatives]. They need to understand [external uncertainty] about [actors/market/technology]. Their choice will change if they observe [two verifiable indicators]”.
Write a KIQ of two sentences at most and three sub-questions. Ask another person to check whether all of them allow for observable evidence.
Map of decisions and intelligence questions
Decision, owner, horizon, alternatives and intelligence question.
See the template →Check it · self-assessment
Check your KIQ
This part needs professional judgement: there is no automatic marking. Complete your work in the practice, rate it with the lesson's rubric and compare it with the indicative solution.
What your answer should include
- The question starts from a decision with a decision maker, alternatives and a horizon.
- The three sub-questions allow for observable evidence.
- You define indicators that would change the choice.
- The decision threshold is proportionate to the cost of acting or waiting, without demanding total certainty.
Indicative solution
Model KIQ: "Which independent technical tests, market conditions and regulatory milestones would justify a limited trial or an alliance by Company A on thermal storage over the next 12 months?". Sub-questions: degree of maturity with reproducible tests; real customer commitments; definitive status of regulatory requirements. Triggers: independent replication of tests; a conditional commercial agreement; formal adoption of a requirement.
Mistake to avoid: Asking "how do we show that we must invest?".
Illustrative solution for a fictional case: reasoned alternatives are valid.
How to recognise a bad question
A KIQ is not useful if nobody can identify an associated decision, if the horizon is unlimited, if it only seeks to confirm a hypothesis, if it aims to describe the whole sector or if it depends on information that cannot be legitimately obtained.
Interactive · Decide
Topic, biased question or KIQ?
Read each formulation and classify it.
Activity: decide whether each formulation is a topic, a biased question or a useful KIQ.
“Monitor artificial intelligence”.
“Why are our competitors going to overtake us?”
“Which independent trial results and which industrial purchasing signals would justify a limited pilot for Company A in the next twelve months?”
“Is thermal storage a technology of the future?”
“How can we show that we should invest?”
Optional extensionDeep dive · The question must also say when to stop investigating
A valuable KIQ does not only open an investigation: it signals when there is enough information to formulate a conditional recommendation. This does not mean reaching absolute certainty. It means identifying an evidence threshold proportionate to the decision and to the cost of delaying it. For a reversible, low-cost trial, a plausible technical signal plus a safety assessment could be enough; for an irreversible investment, additional cross-checks would be needed. The thresholds are agreed with the decision maker; the analyst does not invent them.
Write an assignment sheet with four linked fields: “decision that falls due”, “alternatives”, “evidence that would change the preference” and “what we will still not know”. At Company A, if management has twelve months to decide, a question about regulatory evolution with no deadline can consume the whole project and not settle the choice.
Worked example · From topic to verifiable assignment
Topic: thermal storage. Insufficient question: “Is it a technology of the future?”. Decision: whether or not to devote technical capacity to a limited trial. Revised question: “Which independent test results and which signals of industrial purchasing would justify a limited trial for Company A over the next twelve months?”. Evidence that could change the criterion: reproducible documentation of performance under comparable conditions; actual contracting of a solution by an industrial customer. Blind spot: the real costs of integration into Company A’s processes.
Ask a colleague to take the opposite stance and propose a piece of evidence that favors waiting. If they cannot, the question may have been written to confirm a decision already made.
Acceptance criterion for the exercise
The reader must be able to point out separately who decides, the legitimate alternatives, the horizon, two observable criteria and a piece of information that could refute the initial preference.
Key takeaway: the goal of intelligence is not to monitor more; it is to reduce a specific uncertainty in time to decide better.
References and limits
- Innguma · internal training material. De las necesidades de información al plan de vigilancia (presentación, 2022-05-12) (2022).Information needs and the watch plan.Innguma training material, not public.
- Company A is a fictional case: its data, companies and figures do not describe real markets or organisations.
- The cycle, the critical factors and the intelligence questions are presented as a general Academy method; they do not reproduce the requirements of any standard.
Intelligence Lab · L02
Monitoring plan builder
Ready to apply it? Open the Monitoring plan builder in the Lab and build an editable plan you can download. Output: editable monitoring plan. Works in the browser, no server, no account and no connected AI. The result may need reviewing and adapting to your case.