Krish Ramineni

Krish Ramineni is co-founder and CEO of Fireflies.ai, the meeting assistant that records, transcribes, summarizes, and makes conversations searchable across an organization. He came out of Microsoft product work drowning in calendar invites, pivoted through half a dozen NLP experiments with co-founder Sam Udotong, and landed on voice as the blue-ocean problem email had already solved for text.

Microsoft meetings as the origin pain

He worked as a product manager across Office, Skype, Bing, and NLP teams at Microsoft, including a "customer voice" project that quantified in-app feedback. In a Unite.AI interview he said inspiration for Fireflies had less to do with those products than with meeting load: "Just doing meetings over and over again became my life, and I wanted to change that." Computer science and ML, he added, were means to an end: "I wanted to be able to just press a button and have some sort of magic happen that would impact millions of people."

Five or six pivots into voice

With Sam he hacked through drones, food apps, and crypto wallets in college before a serious NLP bet: auto-generated to-do lists from chat promises on Slack and LinkedIn. Email extensions and chatbots followed. "We pivoted 5-6 times before getting to where we are today," he told Unite.AI. Voice won because you can search old email easily, but "if you had a meeting two hours ago, you probably would have already forgotten everything that was said."

Meetings into a living knowledge base

Fireflies transcribes, summarizes, extracts action items, and can push notes into CRMs and project tools. Ramineni's enterprise pitch is a self-updating knowledge base of customer, candidate, and teammate voice — unlike wikis that go stale. The product goal he states is blunt: "help you review a one-hour call in less than five minutes."

On product-led growth and meeting cost

Growth stayed product-led and freemium: show up on the call, send the recap, let people swipe a card. He frames the mission around wasted meeting spend and unproductive hours, arguing organizations burn enormous time in redundant gatherings that could have been email. The through-line is perfect memory for conversations so humans can stay present and do less admin afterward.