Personal AI assistant for repetitive sales admin (L1)
Give each rep a personal AI assistant (Lindy or Relevance) wired to their Gmail + Calendar. It drafts follow-up emails after meetings, books internal reviews, and chases unanswered threads. Saves 30,60 min/rep/day with no CRM changes required.
The steps
- 01
Pick Lindy or Relevance
Tool: Lindy
Lindy = best for non-technical users, drag-and-drop agent builder, strong Gmail/Calendar/Slack triggers. Relevance AI = more flexible, better for RevOps who want to template the agent and clone it across reps. For an L1 rollout, start with Lindy unless you already have a RevOps developer. Owner: sales ops. Time: 1 hour to evaluate. Pitfall: building from scratch when the tool ships templates for "Sales Assistant", use the template. DoD: one named tool, one assigned admin, billing set up.
- 02
Connect Gmail + Calendar (1 rep first)
Tool: Gmail
Pilot with ONE willing rep. OAuth-connect their Gmail and Google Calendar inside Lindy/Relevance. Scope permissions to read + draft only (do NOT give send-without-confirmation on day 1, trust is fragile). Owner: pilot rep + ops. Time: 10 minutes. Pitfall: granting full send permission immediately and watching the agent email a customer something embarrassing. DoD: pilot rep sees the assistant in their tool tray and can trigger a test draft.
- 03
Wire 3 specific jobs
Tool: Lindy
Don't try to automate everything. Pick 3 high-frequency tasks: 1) After every external meeting (calendar trigger), draft a recap email referencing the meeting transcript (pull from Granola/Fathom) and put it in Gmail Drafts. 2) Every Monday 8am, scan inbox for threads with no reply in 5+ days from prospects and draft a nudge. 3) When a rep gets an email containing "can you do {time}", check calendar and reply with 3 alternatives. Owner: ops + rep co-design each agent. Pitfall: vague instructions like "help with email", agents need specific triggers + specific outputs. DoD: each of the 3 agents fires correctly on test data 3 times in a row.
- 04
Measure & expand to the team
Tool: Manual
After 2 weeks, ask the pilot rep: (a) hours saved per day, (b) one embarrassing thing the agent did, (c) what they'd add. If hours saved > 3/week and no customer-facing incident, clone the setup to 3 more reps. Then the whole team in week 4. Owner: sales manager. Pitfall: rolling out to the whole team in week 1, one bad draft to a customer kills team trust permanently. DoD: a written 1-pager with "hours saved per rep per week" and "3 agents we run" ready to share with leadership.
Tools in this playbook
Next playbooks
Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.
