Lessons from Building AI Products

Inside Gong: How teams work with design partners, their pod structure, autonomy, trust, and more

Topic 3 of 5

Published Jan 2, 2025 · Clip 31:49–38:17 (6:28)

Gong has built AI products longer than most, starting with machine learning before it was mainstream, avoiding early hype pitfalls. Alon advises against extremes: don't over-rely on LLMs for everything or revert to needing data scientists everywhere. Key is retaining AI expertise to guide PMs on feasibility, measure progress (e.g., via Elo ratings), and optimize prompts/models. Examples include deal prediction requiring custom models over LLMs, and fine-tuning for accuracy. Pods now include embedded AI specialists for rapid iteration on LLMs or custom solutions.