A BA I worked with once told me she had run forty stakeholder discovery conversations in her first year and learned almost nothing from them. The questions had been fine. The notes had been thorough. The output documents had checked out with her manager. Yet six months later, when the projects she had scoped were running, she kept finding gaps that the conversations should have caught.
We sat down and looked at how she ran the conversations. The questions were the standard ones. The problem was structural. The opening minute had been wrong, which set the wrong frame, which produced the wrong answers, regardless of how good the questions were.
This piece is about that structure. How a discovery conversation actually works. The opening, the middle, the close, and the five small mistakes that turn it into a status meeting that produces nothing.
The opening minute decides everything
Most BAs open the discovery conversation with some version of “thanks for taking the time, let me share my screen and walk you through what I’m hoping to cover today.” The stakeholder hears: this is a structured interview, I am the subject, the BA has a script.
That framing kills the conversation. The stakeholder shifts into answer-the-questions mode, gives factually correct but flat answers, and the BA leaves with what they came with.
The opening that works is the opposite. No screen-share. No agenda walkthrough. The BA opens with one sentence that gives the stakeholder permission to digress.
“Before we get into anything formal, can I ask what’s been on your mind about this project this week.”
What this does: it signals that you are interested in what they are actually thinking about, not just what they have prepared to say. About seventy percent of the time, the stakeholder will tell you something that wasn’t in any of the prep material. That something is usually the most important information in the conversation.
Three caveats. First, the question has to be genuine. If you ask it as a rhetorical opener and immediately pivot to your real questions, the stakeholder notices and the trust evaporates for the rest of the session. Second, the question has to land in the first sixty seconds. After that the stakeholder has settled into formal mode and you cannot pull them out. Third, you have to be willing to follow the answer for five or ten minutes, even when it seems off-topic, before steering back. Cutting off the digression early defeats the whole purpose.
The middle pivot
After the opening conversation has produced its raw material, the BA has to make a transition that most BA training never describes. The transition from listening to interrogating.
The mistake most BAs make is they never make the transition. They listen the whole time, take notes, leave with raw material, and never push on any of it in the room. This produces sympathetic conversation and bad requirements.
The transition move is small. Somewhere around the fifteen-to-twenty-minute mark, the BA says something like: “I want to push back on something you said earlier, just to make sure I understand it.” Then they pick one of the stakeholder’s claims and challenge it, in a low-stakes way.
For example: “You mentioned the support team is overwhelmed by these tickets. When you say overwhelmed, are we talking about response time, ticket volume, complexity of individual tickets, or something else.” This is not a clarification question. It is a push. The stakeholder has to defend or refine the original claim. The defending or refining is where the real information lives.
Most stakeholders respect this move. It signals that the BA was actually listening rather than waiting for their next scripted question. About twenty percent of stakeholders react badly to it, in which case you have learned something important about who you are working with.
The pivot should happen exactly once. More than once and the conversation becomes adversarial. Zero times and you walked away with status-update notes.
The middle exploration
After the pivot, the conversation moves into its working phase. This is where most BA training focuses, because it is the part of the conversation that looks like work. The questions land here. The notes get taken here. The follow-ups get queued here.
A few mechanics that change the quality of this section.
Match the stakeholder’s specificity. If they say “we need better reporting,” do not ask “what kind of reporting.” Ask “can you walk me through the last time you needed a report and could not get it.” Stakeholders cannot answer abstract questions about reporting. They can answer concrete questions about Tuesday at three pm.
Time-box your own talking. Most BAs talk thirty to forty percent of the conversation. The good ones talk fifteen to twenty. Your job is to listen. Every sentence you say should either be a question or a checked summary, never a contribution.
Ask the same question two different ways. Most stakeholders give you their first-draft answer to any question. The second-draft answer, which is usually more accurate, only comes if you ask the question again with slightly different framing five or ten minutes later. The contradictions between the two answers are diagnostic.
Note what they do not say. Discovery conversations are partly about what the stakeholder says and partly about what they avoid saying. If you ask about the support team’s experience and they describe processes but never mention the support team’s manager, that omission is information.
The close
The closing minute matters as much as the opening minute and gets even less attention.
The standard close — “thanks so much, I’ll send you the meeting notes and circle back if I have follow-ups” — produces a useful artifact (the notes) but ends the conversation flat. The stakeholder leaves having performed a task. They do not leave thinking about anything new.
The close that works ends with one question the stakeholder has to think about after the meeting. Not a question they answer in the room. A question they take with them.
“Before we wrap, here’s one thing I’d love your read on, but not now. What would have to be true for this project to succeed in the way you actually want it to succeed, not just hit the timeline.”
The stakeholder will not answer in the room, because the question requires real thinking. They will, in many cases, send you a follow-up email three or four days later with a longer and more interesting answer than you would have gotten in the meeting. That follow-up is usually the most valuable artifact of the entire discovery cycle.
Five mistakes that turn it into a status meeting
To close, the failure modes worth naming explicitly.
One. Opening with “let me share my screen.” The screen-share frames the conversation as a presentation. Don’t.
Two. Reading questions from a list. The list is for your prep, not your delivery. Memorise the three or four highest-priority questions and let the conversation route to them naturally.
Three. Filling silence. After a stakeholder finishes a thought, most BAs fill the next two seconds with a follow-up. The two seconds of silence are exactly when the stakeholder adds the most useful sentence they had not planned to say.
Four. Confirming too quickly. If the stakeholder says “yes, exactly,” and you move on, you have just installed a misunderstanding into the requirements document. Ask “can you say more about what exactly looks like for you” before accepting any “yes.”
Five. Skipping the close question. The biggest difference between a discovery conversation that produces a real requirement and one that produces a status update is whether the stakeholder leaves thinking about the project differently than they came in. The close question is what makes that happen.
What this is really about
The structure above is a way of designing a conversation so that the stakeholder ends up doing the analytical work in their own head, with you holding space for them to do it. The BA’s job in a discovery conversation is not to extract requirements. It is to create the conditions in which the stakeholder, often for the first time, sees their own problem clearly. The requirements emerge from the seeing.
The BAs who advance into senior roles all run conversations like this, even if they cannot articulate why. The ones who plateau run conversations as structured interviews, get factually correct but flat answers, and write requirements documents that pass review and fail in production.
The structure is small. The questions are not the point. The opening minute, the middle pivot, the silences, the close question. That is the whole craft, in one hour.
The next piece moves from the conversation to what to write up after it — the format of the discovery artifact, how detailed it should be, and what should never be written down. Subscribe below to get it.