Ask Slackbot is a monthly column that answers reader questions about work in the age of agentic AI. Answered by the Slackbot of someone in a different Salesforce role each month, it’s grounded in that employee’s real Slack conversations, customer implementations, research, and relevant documents. It can search messages, read files, synthesize patterns, and connect insights — all while respecting permissions and data privacy within the Salesforce Trust Layer. It offers some broad life and work lessons, too. Got a question particular to your job function? Drop us a line.
Dear Slackbot,
I’m a forward deployed engineer preparing for a discovery call with a new customer who already has firm opinions about what they want built. I’m expected to be both the technical expert and an attentive listener. I tend to fall into one of two patterns: either I accept their first answer without probing or I question things so much that I feel like I’m interrogating them. How can I challenge their assumptions and find the underlying problem without coming across as adversarial?
— Expert-Or-Listener-Pick-One
Dear Expert-Or-Listener-Pick-One,
You don’t have to pick one. A curious interlocutor can be both the smartest person in the room and the most interested person in the room at the same time.
Customers who arrive with firm opinions have usually lived with the problem longer than you have. Their solution is a compressed story about a pain point, and your job is to unpack it. Don’t dispute the answer. Ask what happened that makes their approach seem like the obvious answer. In a recent engagement you had, I noticed that the original approach was replaced mid-build by a different design. The early ask was reasonable. It just wasn’t the whole picture.
“Why do you want that particular feature?” sounds like a challenge. “What does that feature unlock for your team?” sounds like interest. You get the same information from the customer, but you don’t ruffle any feathers. In one exchange I came across, you asked a single clarifying question: “We tend to prefer using these existing components, but tell me what you get if we pursue a newer design?” That got you good intel without implying the customer was taking a wrong turn.
The data says what it says. “Here’s what I’m seeing in the numbers” is a much friendlier sentence than “I disagree.” In one message exchange I read, you described the customer’s existing setup — hundreds of mapping records, dozens of categories — in concrete numbers before proposing anything. Numbers can let the customer reach a conclusion on their own, and that’s more likely to stick.
Nothing disarms a room like an expert admitting a gap. I noticed you’ve opened technical write-ups with statements like “here’s what I’m confident about” and “here’s what I’d want to validate first” before laying out options. Socrates himself said, “I know that I know nothing.” You’re in good company.
Leave the call with fewer answers than you wanted and better questions than you brought. You can build the rest later.
— Slackbot
Your 3 next best steps ✅
- Write three “unpack” questions for each firm opinion the customer has. Phrase them as “What happens when …?” and “What does that unlock …?” not “Why?” Pick the best one per topic and keep the rest in your back pocket.
- Pull one concrete example from a past engagement where the first ask and the final build diverged. Keep it anonymous, and share: “On a similar project, we found …” Your probing questions will read as experience, not a direct challenge.
- Set a pacing rule. After every two questions, reflect back what you heard. That can turn a potentially tense interrogation into a productive discovery call.
How Slackbot generated this column
To answer this question, Slackbot read the asker’s question and turned it into a column-ready dilemma about balancing expertise with listening. It scanned the asker’s recent engagement channels, threads, and documents for moments where an initial customer ask changed during a project, where clarifying questions were used, and where data was brought to the conversation. Every customer, colleague, project, and channel name was removed, and the signals were reduced to general patterns. Slackbot then applied editorial judgment to match the Ask Slackbot voice.
Stealable prompt (forward deployed engineer):
Look at my last five customer engagement threads and kickoff notes. For each, tell me what the customer first asked for, what the project ended up needing, and the moment the difference first showed up. Then draft five open-ended discovery questions for my next kickoff that would have surfaced that gap earlier. Keep the questions curious, not challenging.
