BA · 8 chapters · CHAPTER 04

Elicitation: How to Discover What People Actually Need

Interviews, workshops, observation and the art of asking better questions

Chapter 04 of 8Read as a practical lesson

Elicitation is discovery, not transcription

Stakeholders rarely arrive with perfectly structured requirements. They describe symptoms, preferences, workarounds, frustrations and desired features. The BA's job is to turn those conversations into shared understanding.

Prepare before the meeting

Know what you already understand, what remains uncertain, who needs to be involved, and what decision the session should enable.

01Prepare
02Explore
03Probe
04Reflect back
05Document
06Confirm

Question patterns that work

QuestionPractical answer
Open“Walk me through what happens today.”
Why“What makes that step necessary?”
Exception“What happens when this does not go as expected?”
Evidence“How do you know this is a problem?”
Outcome“What would be different if this worked well?”
Constraint“What cannot change?”

Choose the technique to fit the problem

Use interviews when you need depth, workshops when you need alignment, observation when people struggle to describe real behavior, surveys when you need breadth, and prototypes when a visual conversation is faster than abstract discussion.

Listen for hidden requirements

“We need this report every morning” might hide a scheduling requirement. “Only managers should see this” might hide an authorization requirement. “It must never fail” might hide an availability or recovery expectation.

✓Record decisions, not just opinions.
✓Capture assumptions separately.
✓Ask for examples and exceptions.
✓Repeat important statements in your own words.
✓Confirm what is still unknown.

After the meeting

Document the result quickly while context is fresh. Then confirm the interpretation with the people who supplied the information. Confirmation is part of elicitation, not administrative cleanup.

Elicitation is where good questions pay for themselves

Many requirements are hidden inside stories. A user says, “We normally refund it, except when the item is damaged.” That one sentence contains a process rule, an exception and an implied decision authority. Slow down and unpack it.

Use the current process as your anchor

When someone describes the desired future state, bring the discussion back to today. Ask what triggers the process, what information enters it, who makes each decision, what system is used, and where work waits. The contrast between As-Is and To-Be gives you the material for a meaningful gap analysis.

Interview technique

Instead of asking, “What fields should the form have?”, ask, “Show me the last time you completed this form. What information did you need at each step, and where did you have to look it up?” The second question exposes actual behavior rather than remembered ideal behavior.

Close the loop in the same meeting

Near the end, replay your understanding: “So the normal path is A → B → C, but when condition X occurs, the order goes to Y. Finance owns the final approval. Is that accurate?” This gives stakeholders a chance to correct your model while the details are still fresh.

Practice this

Run a mock discovery for an online return process. Start with “walk me through the last return,” then ask follow-up questions until you can describe the happy path, two exceptions, the business rule behind each exception and who owns each decision.