Running a discovery workshop for your own team is hard. Running one for a client is harder in a completely different way. The facilitation challenges are structural, not just interpersonal. You're trying to extract honest signal from a group of people who are also evaluating you, whose internal politics you don't fully understand, and who will sometimes tell you what they think you want to hear rather than what they actually believe.
Agencies that do this well have learned a few things the hard way. The rest of this post is those things.
The authority problem
When you facilitate an internal workshop, the facilitator's authority is positional. Everyone knows who's running it and why. With a client group, your authority is provisional - it exists only to the extent that the participants choose to grant it. The moment the most senior person in the room starts treating the session as a presentation opportunity rather than a working session, you've lost the room.
The solution is pre-work, not confrontation. Before the workshop, you need to brief the senior stakeholder privately. Not to tell them how to behave, but to align on what success looks like. "At the end of this session, we want every person to have surfaced at least one thing they believe is true about your customers that they haven't said publicly before." When the senior person understands that the goal is signal, not performance, they're more likely to create space for others rather than fill it.
This is one of the most important things you can do before the room fills up. The work of facilitation happens before the session starts.
The hierarchy problem
Client workshops almost always have visible hierarchy. The VP is present. So is the junior researcher who knows the most about the customer but has learned not to contradict the VP. The output of a workshop like this isn't discovery - it's a polished version of whatever the VP already believes, with everyone else nodding along.
The structural solution is anonymity at the right moments. Not for everything - you want named ownership of ideas when building a roadmap, because accountability matters. But for idea generation and prioritization, anonymity removes the most distorting variable in the room. When nobody knows whose idea is whose, the VP's suggestion competes on equal footing with the researcher's.
This isn't a soft facilitation trick. It's a design decision about the process. Tools that let participants submit ideas or cast votes without their names attached don't make the session feel less professional - they make the output more honest.
The interpretation problem
Here's something agencies almost never admit: you often understand the client's customer better than the client does. Not because you're smarter, but because you have outside perspective and you've done this with other companies. The client has been inside their own assumptions for so long that they can't see them.
The temptation is to tell the client what you've observed. Don't. Or at least, not directly. If you tell a room of senior people that their product assumptions are wrong, they'll defend those assumptions. If you run a process that surfaces the contradiction internally - where the data or the customer voice is in the room, not just your interpretation of it - the client reaches the insight themselves. That insight sticks. Your pronouncement doesn't.
This means building time into the workshop agenda for contact with actual customer language. Quotes, call recordings, survey responses, support tickets. Not your synthesis of them - the raw material. When the VP reads a support ticket that directly contradicts their stated belief about what customers want, the room changes. That's the moment discovery actually happens.
The deliverable problem
Clients hire agencies in part because they expect a document at the end. The workshop is framed as the means; the report is the end. This creates a perverse incentive: the facilitator is tempted to run a session that produces clean, documentable outputs, rather than one that surfaces messy, honest signal.
The way out of this is to reframe the deliverable contract before the session, not after. The output of a discovery workshop is not a report. The output is a shared set of decisions your team now has permission to make. The report is just the record of how you got there. This reframe matters because it changes what the session is optimized for. If the client understands that the goal is decision-clarity, not content production, you have permission to sit with productive tension longer - to not resolve things artificially just because the clock says it's time to synthesize.
What good structure actually gives you
The agencies that run great client workshops aren't better at reading the room. They're better at designing the room before anyone walks in. They use structured formats for idea generation so that introverts produce as much as extroverts. They use anonymous input at the prioritization stage so that hierarchy doesn't silently decide the outcome. They brief sponsors in advance to prevent the session from becoming a presentation.
Structure isn't the opposite of creativity. It's what makes creativity safe in a room where the stakes are high and the trust is new.
Go deeper
If you run workshops for clients, Bandos was built with your workflow in mind. How Bandos works with agencies covers the specific tools for managing client sessions - from anonymous ideation to structured voting - without requiring your clients to learn a new platform. When your client's VP and junior researcher need to contribute equally, anonymous voting is the mechanism that makes that possible without anyone losing face.