A workshop can feel productive without changing anything. There are sticky notes, energetic conversations and a photograph of the whiteboard. Two weeks later, the designer and developer still disagree about what the product is supposed to do.

The value of discovery is in the decisions it makes possible. Before scheduling a session, agree on the questions the team needs answered and the evidence it needs to answer them.

Bring a real workflow into the room

“We need a customer portal” is a proposed answer. Start a little earlier. What are customers trying to do today? Where do they wait? What information does an employee copy from one place to another? Which exceptions cause the most confusion?

Invite someone who performs that work, and ask them to demonstrate an ordinary case. Then ask for a difficult one. A late payment, an incomplete application or a booking change often exposes requirements that an ideal journey hides.

Use actual examples with sensitive details removed. A sample invoice or an anonymized support request gives everyone something concrete to discuss. It is harder to resolve a disagreement when each person is imagining a different customer.

Separate decisions from assumptions

Keep a visible record with three columns: agreed, assumed and unknown. “Managers approve refunds” might be agreed. “Most customers want to use a mobile app” may be an assumption. “Who approves a manager's own refund?” could remain unknown until the finance lead joins.

This distinction protects the schedule. An assumption should not quietly turn into a development requirement simply because nobody challenged it in the meeting. Assign an owner and a date to important unknowns, especially those involving permissions, payments or external systems.

Describe a first release in plain language

At the end of the session, someone should be able to describe the first release without pointing at a diagram. For example: customers submit a request, staff check the information, a manager approves it, and the customer receives a status update. List the exceptions that must work on day one.

Then name the work that can wait. A separate later list prevents useful future ideas from crowding the first delivery. It also lets stakeholders see that an idea has been recorded rather than forgotten.

Leave with a small set of usable documents

  • A one-page problem statement with the intended users and outcome.
  • A workflow showing the main path and important exceptions.
  • A first-release scope with explicit exclusions.
  • A list of risks, unanswered questions and responsible people.
  • Acceptance examples that explain what successful behavior looks like.

A polished slide deck is optional. Shared understanding is essential. Ask someone who missed the workshop to read the outputs and explain the proposed product back to you. Their questions will show where the team still relies on conversation that never made it onto the page.