Teams Insider

Telcos Moving Into AI With Microsoft Teams and Copilot With Rick Garcia

Rick Garcia, EVP of Product and Marketing at Momentum, explains how a telco with 7,000+ mid-market and enterprise customers is building an AI practice on top of its Teams Phone and connectivity business. He covers packaging repeatable workflows such as missed-call-to-SMS, borrowing ideas from n8n an

Prefer audio?

Who should listen: Telco and UCaaS product leaders, Microsoft partners building AI propositions on Teams Phone, and M365 service owners weighing Power Automate and Azure OpenAI workflows against third-party automation platforms.

Guest: Rick Garcia — EVP Product & Marketing, Momentum

Many thanks to Momentum, sponsor of this episode.

Key insights

Telcos already own the 'plumbing' — bandwidth, dial tone, Teams Phone — so the differentiator is shifting from selling seats and lines to acting as a business consultant that delivers workflows on top of what is already installed. ▶ 9:40
Carriers cannot copy the traditional Microsoft partner model of month-long on-site professional services. Momentum instead hunts for simple, repeatable workflow packages it can sell across thousands of customers without getting stuck in a year-long bespoke build. ▶ 6:04
Garcia cites failure rates of 75–95% for AI pilots, and attributes them to buyers expecting an AI platform to replace an entire customer service team rather than solving a defined problem. ▶ 10:10
There is no missed-call endpoint in Graph, so the team built a Power Automate flow that polls Graph for calls in the last 5–10 minutes that were potentially missed, then triggers an AI-generated text response — a reusable building block for any business where a missed call equals lost revenue. ▶ 7:06
AI-assisted texting sells because both staff and customers already understand the medium; the change is replacing static keyword routing ('text back YES/NO/SCHEDULE') with an interpretation layer that can handle 'I have a problem with my billing'. ▶ 17:21
A practical measure of internal adoption: Garcia watches the demo tenant to see sales people building their own AI texting agents and testing use cases like 'I'd like to schedule a test drive' — visibility he cannot get as easily on the Teams Phone side. ▶ 18:55
Most publicly shared n8n and Pipedream automations are aimed at micro-businesses and hang off Gmail; the value for enterprise partners is translating those patterns into Outlook, Power Automate and Azure, then waiting for Microsoft to ship the equivalent natively. ▶ 21:30
The lag between a frontier model or protocol landing and it appearing in the Microsoft stack is roughly six months and shrinking — GPT-5 arrived same-day, and MCP went Anthropic, then OpenAI, then Microsoft. ▶ 22:32
The next step beyond workflows is contextual insight rather than BI: call duration and answer-time stats mean little to a CEO, whereas combining call, email, chat and SharePoint data to flag five accounts likely to churn on negative sentiment is a business decision. ▶ 23:32