Teams Insider

Inside Gong’s AI Microsoft Teams Integration

Patrick Burke leads go-to-market for technology partnerships at Gong, the revenue AI platform. He explains how Gong captures and analyses sales conversations, why it partners with Numonix's recording-as-a-service to get compliant media from Microsoft Teams calls, and where agentic AI and MCP could t

Prefer audio?

Who should listen: Microsoft 365 service owners and UC professionals who are being asked to approve or integrate third-party meeting-insight and revenue AI tools, and anyone weighing certified compliance recording engines against building on Teams media APIs directly.

Guest: Patrick Burke — Go-to-market lead for technology partnerships, Gong

Many thanks to Numonix, sponsor of this episode.

Key insights

Gong began as a conversational intelligence product after its founder missed a quarter's forecast with no visibility into why; it now spans three products — Gong Enable (revenue/conversational intelligence), Gong Engage (sales engagement) and Gong Forecast (deal boards and forecasting). ▶ 4:08
The core data-loss problem Burke quotes: roughly 5,000 words are spoken on a sales call and, on average, about 50 make it back into the CRM manually — and 50 is a good rep. ▶ 5:08
Native CRM integrations are limited to Dynamics, Salesforce and HubSpot; on the UC side Gong supports Teams, Zoom and Google Meet, with custom integrations possible beyond that. ▶ 7:12
Burke concedes call recording and transcription are now commoditised — the differentiation is the recently released agent layer that can query a single call, all calls with an account, or the entirety of activity with that account, out of the box or with custom extractions. ▶ 8:12
Gong claims around 40 of its own native LLMs, built because transcribing and understanding call context was never something it could do manually. ▶ 6:42
Rather than building on Microsoft's media APIs directly, Gong routes Teams capture through Numonix's certified compliance recording engine (their TRaaS service) — covering desktop, web and mobile scenarios and removing the need for enterprise customers to re-scrutinise compliance. Arbuthnot's point: you can DIY with the APIs, but if Microsoft changes something and you are not certified, there is nobody to turn to. ▶ 14:51
Teams mentions in Gong's own customer call data have gone up and to the right every quarter for five years, with the current quarter the highest yet — Burke uses keyword mention counts quarter over quarter as a market gut check. ▶ 13:50
On MCP, Burke sees Gong acting as both server and host — hub and spoke — so sales data can be queried from ChatGPT, Copilot or Claude and pushed back into Gong; he stresses this is not a roadmap commitment. ▶ 17:28
A Gong Labs finding from analysing call data: cursing more on calls correlates with better deal outcomes, as the buyer perceives the rep as more personable. ▶ 12:18