How BioRender cut sales follow-up drafting from 20 minutes to about 3 with one Zap

The best part of this build is not the AI. It is where the AI stops.

What happened

Zapier named Jocelyne Mendez-Guzman of BioRender's RevOps team its 2026 Sales AI Builder of the Year for a follow-up system that runs after every logged sales call. Zapier's write-up describes five steps:

1. Capture: a Gong webhook sends the transcript and call metadata to Zapier.
2. Structure: JavaScript cleans Gong's output into key and value fields per participant.
3. Interpret: AI by Zapier extracts the account, recipient, key points and next steps.
4. Route: the workflow identifies each internal participant's role and the call type, then picks one of six paths covering business development, account executives and customer success.
5. Assemble: code stitches those details into a role-specific template.

The draft is formatted for Gmail, logged in Google Sheets and sent to the rep in Slack with a button to open it. Zapier's winners page says 42 reps and CSMs have used 3,468 drafts, with follow-up time dropping from 15 to 20 minutes to about three. It was rolled out team by team, with demo videos and hands-on training.

My take

This is the pattern I push clients toward, and it is nice to see it win. Three details matter:

• AI handles the fuzzy part, reading a messy conversation. Deterministic code handles the exact part, like pricing templates, so nothing gets invented.
• The rep stays in control. The output is a draft in Gmail, not an auto-sent email, which is why people trust it.
• Adoption was designed, not assumed. Staged rollout plus training beats a big launch nobody uses.

Zapier's own takeaway says it well: measure adoption, not output. If your team uses HubSpot, GoHighLevel or Pipedrive instead of Gong, the same shape works: a call or meeting event triggers extraction, routing picks the template, and the rep approves the send. Start with one team and one call type, then expand.