Gong

From listening to a few calls to recording every sales call

A large sales organization holds thousands of calls a month, and a human manager has time to listen to very few of them. Everything else (what was actually said, which objections kept coming up, where a deal started to stall) stays out of the picture, and the forecast relied mainly on what the salespeople reported themselves. Gong, founded in August 2015 by Amit Bendov and Eilon Reshef, built a product that records and transcribes the calls, emails and meetings of the whole organization and runs language analysis on them. A competitor, Chorus.ai, also worked in the same field, analyzing sales calls with AI; in July 2021 ZoomInfo announced it was buying it for $575 million.

Two outputs: personal coaching and revenue forecasting

Two outputs follow from this. The first is coaching: the system identifies what successful salespeople do differently from the others (how much they talk versus listen, when they bring up price) and produces personal feedback from it. Second Nature offers a different approach to coaching: salespeople practice in simulated calls with AI characters, and the system gives them a score and feedback. The second output is forecasting: instead of relying on a subjective estimate, the model flags deals whose call signals do not match what was reported about them. The category the company helped define is called "revenue intelligence."

Funding path and growth figures up to $7.25 billion

The financial path stands out: $65 million at a $750 million valuation (December 2019), $200 million at $2.2 billion (August 2020), and $250 million at a valuation of about $7.25 billion in June 2021, one of the highest valuations ever for a private Israeli company. The June 2021 round (Series E) was led by Franklin Templeton, with participation from Coatue, Salesforce Ventures, Sequoia, Thrive Capital and Tiger Global, bringing Gong's total funding at the time to $584 million. In the official announcement of the round, the company cited 2.3x growth in annual recurring revenue (ARR) between the first quarter of 2020 and the first quarter of 2021, more than 2,000 customers (including three Fortune 20 companies that joined close to the announcement), about 550 employees worldwide, and the capture of more than half a billion customer interactions since its founding. "As professionals across the globe were forced to connect with customers remotely for the first time, they became even more reliant on data and insights to win the hearts and minds of their customers," Gong CEO Amit Bendov said at the time. In early 2025 the company reported passing $300 million in annual recurring revenue, and as of 2026 it employs about 1,200 people, about a third of them in Israel.

The privacy question of systematic recording

The field also raises a built-in question: systematically recording every call is employee monitoring, and the rules on consent to recording differ from country to country.

From a model advantage to a data advantage: what really protects Gong

On the technical side, the company built much of its capability before the era of large language models, on transcription and classification developed specifically for sales. The availability of powerful general models changed the balance of power: what was once a rare technological advantage became relatively accessible, and the advantage moved to the historical data that had accumulated and to deep integration into the customer's work systems.

This is a pattern that repeats across the whole category: what protects a mature AI company is not the model but the data only it sees, and the customer's cost of switching to a competitor. A company that holds years of labeled sales calls with their outcomes can train on something a new competitor simply does not have, even if the competitor's model is just as good.