AI Sales Pilots: From Tech Demos to Metric-Driven Wins
🎧 PodShort
36 min squeezed to 2
ARMARM Sales Tech New

Sean Frank
Founder and CEO at Gong
Mark
Unknown at Unknown
Full episode from 30 Minutes to President's Club
Quotable Moments
And that is sort of the reverse engineering of the value case. It becomes a natural next step because I've already shown how I can influence the metrics over a four-week period.
Engineers respect other engineers. Why? Because they're doing the work.
Key Insights
- The shift in product development for AI in sales is moving from demonstrating what's possible to actively showing how a specific metric can be improved within a short pilot period.
- Success criteria for pilots are now about moving a specific metric that the buyer cares about, rather than just showcasing the technology's capabilities.
- To get buyers to provide baseline metrics, sellers need to demonstrate how the pilot will tangibly improve those numbers within the pilot's timeframe.
- Effective sales pilots focus on moving leading indicators that are within the seller's control, rather than solely on lagging indicators like revenue which have more variables.
- Knowledge coverage (having a documented answer to an incoming question) is a key metric that AI can significantly improve, often by threefold during a pilot.
- The distinction between SaaS and AI in the current market lies in the 'build vs. buy' mentality; AI solutions with embedded 'build' capabilities through an FE offer a more collaborative and integrated experience.
- A successful pilot requires aligning on success criteria upfront, building a 'build team' with technical expertise, and ensuring the pilot demonstrably moves a key metric.
- Common pilot mistakes include not aligning on success criteria, not building a technical champion on the buyer's side, and not mapping out the tangible value delivered by the end of the pilot.
Metrics Mentioned
- $5 million (Annual cost a company was paying an outsourcer per ticket, with the goal of reducing it by 30%.)
- 30% (The goal for cost reduction per ticket from an outsourcer.)
- 3-fold (The typical improvement in knowledge base metrics during a pilot.)
- 4 weeks (The duration of a typical pilot.)
- 10,000 (The cost of a 4-week pilot, which includes the value of a fully updated knowledge base.)
RevBots.ai View:
- AI Sprinkler teams bolt on pilots without transforming processes, risking misaligned metrics.
- ARM maturity requires embedding AI to move beyond pilots into continuous metric improvement.
- SaaS Hoarder teams struggle with 'build vs. buy' decisions, delaying AI adoption.
- Tab Hopper teams lack technical champions, making AI pilots ineffective.
Join The RevBots ARMy
The insider daily for Autonomous Revenue Masters.