The problem Sales reps were spending significant time searching for content and answers during live customer conversations, and support teams were fielding repetitive enablement questions that didn't require human intervention.
Context and constraints We needed a solution that could integrate with the client’s existing content infrastructure, work within enterprise security requirements, and be shipped fast enough to influence renewal conversations with key accounts. Engineering bandwidth was limited and the roadmap was already crowded.
Process I ran discovery across three user segments -- AEs, SEs, and CS reps -- to map where friction was highest and where AI assistance would generate the most immediate value. I scoped an MVP focused on in-app Q&A over a curated content corpus, deprioritizing generative features until retrieval quality was validated. I owned the roadmap, coordinated across engineering and design, and partnered with CS and Sales to define the GTM rollout sequence.
Decision Shipped retrieval-augmented Q&A as the core feature, with a feedback loop built in from day one to surface low-confidence answers for human review. Deferred voice and proactive nudge features to phase 2.
Outcome 55% reduction in support ticket volume. 73% increase in closed sales among accounts using the copilot in late-stage conversations. Feature became a centerpiece of the ACV expansion pitch.
Made with Bullet