Type 3-word notes while the prospect talks. Claude outputs the reflection line, 3 use cases, and the close script — ready to read back before the call ends.
Discovery calls are 30 minutes of listening and synthesis — and then you have to close on the spot with specific use cases. That's hard to do in your head while also running the conversation.
Write down what the prospect said. Try to capture the important stuff without losing the thread of the conversation.
Transition between structured questions while processing what they just said. Easy to skip one or ask them out of order when you're focused on listening.
Based on everything they said, come up with 3 specific ways AI could help their business — right now, out loud, while they're watching you think.
Try to connect what they said to the offer without a script. Either pitch too hard or leave money on the table — rarely hit the right note without prep.
After the call ends, write the follow-up email from memory. 20 minutes of post-call work to capture what happened in 30 minutes of conversation.
"I was so focused on listening that I had nothing prepared to say when it was time to close."
The /discovery skill runs in Claude Code beside the Meet window. I type short notes as they talk. Claude synthesizes everything by the time I need to close.
Run /discovery in Claude Code. Copilot is live beside the Meet window
As they answer each question, type shorthand: "no systems", "manual invoices", "scaling"
Generates a reflection of their situation using their exact language
Specific to their business, their tools, their bottleneck
Read it back verbatim or adapt it. Client folder and CRM updated automatically
A 2-sentence summary of their situation in their words. Read it back and watch them say "exactly."
Not generic AI pitches — specific applications built from what they actually told you.
Word-for-word language connecting their situation to the OPSYS offer. Nothing improvised.
Three honest cards from the build — what worked, what surprised me, what I'd change.
The bottleneck wasn't the discovery questions — those I could ask. The bottleneck was synthesis: turning 30 minutes of notes into a specific, persuasive use-case pitch in the last 5 minutes of the call. I realized I was asking Claude to do this after every call anyway — so I moved it into the call.
I expected the use cases to be the closer. They're not. When you read back someone's own situation in their own words, they feel understood — and that's what converts. The use cases confirm the fit. But the reflection line creates the trust that makes someone say yes.
Right now I capture notes during the call. What I want is for Fathom to record the transcript and auto-push it to the skill after — so the CRM entry is written by AI, not me. Zero post-call work. Everything logged before I close the laptop.
The call is covered. Pre-call prep and post-call admin are next.
Right now I type notes during the call. The next version pulls the Fathom transcript after the call ends and writes the CRM entry automatically — prospect name, pain points, use cases discussed, next step. Zero post-call admin. Everything logged before I close the laptop.
30 minutes before every discovery call, a brief fires — LinkedIn profile summary, company context, likely pain points based on their industry. Walk into every call already knowing more than they expect. The reflection lands harder when you already know their world.
Right now each call starts fresh. The next version tracks what objections came up, what use cases landed, what language converted. Over time the copilot gets sharper — it knows what works for this type of prospect before the call starts.
Notes in, reflection out. The close script ready before the call ends.
Demo recording coming soon.
Check back in a few days.
Book a 30-minute discovery call. You'll experience it firsthand — and see what it could look like for your business.
Book a discovery call → See more projects