Melissa
Gordon

The sport of women's pole vault wasn't even added to the Olympic program until Sydney 2000. By then, Melissa Gordon was already a two-time NCAA Division I All-American at Stanford, competing at a level the sport's governing bodies were still debating whether to recognize. Three national high school championships. A career built around running full speed at something most people were still deciding was worth clearing.

That same instinct - spotting the bar before the field does, committing to the approach when others are still deciding - has defined the arc of a career that has moved from Oracle's Northeast enterprise accounts to the frontier of generative AI. Gordon joined Rasa as CEO in February 2022, stepping into one of the more precisely positioned bets in enterprise software: that the future of AI in regulated industries isn't about chatbots that surprise you, but ones you can actually control.

Rasa was already a force by the time Gordon arrived. Founded in Berlin in 2016 by Alan Nichol and Alex Weidauer, the platform had accumulated 50 million developer downloads and a devoted open-source community. What it needed was a CEO who knew how to turn developer enthusiasm into enterprise revenue - someone who had spent time inside Oracle's F500 accounts group, who had scaled sales at MemSQL before it became SingleStore, who understood why the procurement cycle at a major bank looks nothing like a GitHub star count.

Gordon, who spent nearly a decade at Oracle and led worldwide sales and customer success at SingleStore (then MemSQL), had built exactly that muscle. Her tenure at Tradeshift as SVP & GM Marketplace added supply chain complexity to the resume. By the time she landed at Rasa, she understood enterprise software not as a category, but as a set of human relationships with compliance departments, IT committees, and CFOs who have been burned before.

The pivot she orchestrated was ambitious. Generative AI arrived and threatened to make Rasa's existing NLU-based approach look dated. Gordon's response wasn't to pivot away from control - it was to build control into the new paradigm. The result was CALM: Conversational AI with Language Models. A framework that uses LLMs for language understanding while keeping business logic deterministic. It doesn't guess what your bank wants to say to a customer. It knows, because you told it.

The positioning landed. Rasa doubled its annual recurring revenue in 2023. February 2024 brought a $30 million Series C co-led by StepStone Group and PayPal Ventures, with Andreessen Horowitz, Accel, and Basis Set Ventures continuing their backing. The round valued the company on an upward trajectory from a prior $26M Series B. American Express, Deutsche Telekom, and two of the world's three largest banks are now on the client roster.

The client list isn't a vanity metric. It's a signal about what Rasa actually sells. In regulated industries - banking, telecom, healthcare - AI adoption isn't blocked by skepticism about the technology. It's blocked by questions about accountability. Who owns the data? What happens if the model says the wrong thing? Can we audit the decision trail? Rasa's architecture - on-premises deployment, deterministic business logic, model customization without vendor lock-in - is designed to answer all three.

In 2025, Gordon moved the company further into voice. A strategic partnership with Cartesia brought Sonic, Cartesia's real-time voice synthesis built on state space models, into the Rasa platform. The combination: enterprise-grade conversational logic with low-latency, hyper-realistic voice output. Target markets are banking, telecom, and healthcare. Deployment options include on-premises, for the institutions that won't put customer conversations through a third-party cloud.

Gordon presented the company's vision at the Montgomery Summit in 2024, where Emmy-nominated journalist Jacki Karsh interviewed her on building high-performance teams and the future of enterprise AI. The through line between the interview and the product roadmap is not hard to spot: both are about knowing when to commit to the approach and trusting the system you've built to carry you over.

A Harvard Business School executive education program in Authentic Leadership Development rounds out the academic formation. The title is worth noting: not leadership, but authentic leadership. For someone who spent her athletic career in a discipline the establishment was still deciding whether to legitimize, authenticity isn't a management buzzword. It's a survival skill.

What Gordon has built at Rasa is a bet that enterprises don't need AI that's impressive - they need AI that's correct. Not chatbots that can write a poem, but assistants that won't give a bank customer wrong information about their mortgage. In an era of generative AI hype, that's a contrarian position. It also happens to be the one that lands $30M from some of the sharpest checks in enterprise software and the business of two out of three of the world's biggest banks.

The pole vault bar doesn't care about your credentials. It falls if you miss and holds if you clear. Melissa Gordon has spent her career building organizations around that same honest feedback loop. At Rasa, the bar is nothing less than enterprise AI that actually works - not in the demo, but in production, at scale, in industries where mistakes cost more than reputation.