In late 2022, while large language models were still a curio for researchers, Shay Levi looked at the CIO landscape and saw a traffic jam forming two years out.
The idea arrived before Unframe existed. Levi was still co-founder and CTO at Noname Security, a company he had helped build from a whiteboard sketch into one of the fastest-growing API security platforms in the world - 250 people, $40M ARR, a trajectory that would end in a roughly $500 million acquisition by Akamai. But a different problem had lodged itself in his thinking.
He describes what he saw plainly: "I basically recognized that enterprises will definitely want to adopt AI, but I kind of foresaw that they were going to struggle." The diagnosis wasn't complicated. Enterprise software is rigid. AI is fast. The collision was obvious to anyone paying close attention to both sides simultaneously. Levi was paying very close attention.
He waited. He watched. He expected someone to build the solution he was already sketching in his head. By late 2023, it was clear that no one was building it at the speed and scale the market needed. He left Noname - before the Akamai acquisition discussions even started - and in January 2024 co-founded Unframe with two colleagues who had walked the same halls at Noname: Larissa Schneider, who became COO, and Adi Azarya, who took VP of R&D.
The founding team is not a coincidence. Three people who had already built something hard together, shipping quickly under pressure, choosing to bet on each other again. That pattern - trust built through shared scar tissue - shows up in how Unframe operates.
Enterprises didn't just need access to AI. They needed a platform that gave them control, speed, and flexibility at the same time.
- Shay Levi, Co-Founder & CEO, Unframe
The Founding Insight
95% of independent enterprise AI projects fail. Unframe was built specifically to fix that number - by treating AI deployment as an infrastructure problem, not a software problem.
Unframe emerged from stealth on April 3, 2025, announcing a $50M Series A backed by Bessemer Venture Partners, TLV Partners, Craft Ventures, Third Point Ventures, SentinelOne Ventures, Cerca Partners, and Terra Nova Ventures. The announcement dropped alongside something more compelling than a press release: $100M in total contract value, booked in the company's first year of selling. By May 2026, a $50M Series B led by Highland Europe closed, and Calcalist ranked Unframe #2 on its list of Israel's 50 most promising startups.
The metric everyone keeps pointing to is the 400% net revenue retention. That number means existing Unframe customers are, on average, spending four times what they started with. In enterprise SaaS, NRR above 120% is considered excellent. 400% is a different category of signal - it suggests that customers aren't just renewing, they are expanding aggressively because the product keeps finding new problems to solve.
The product itself is built around a simple premise: stop asking enterprises to figure out AI and start delivering working outcomes. Unframe provides AI agents, data abstraction, observability, and modernization capabilities packaged as a complete platform. The company calls its delivery model "Blueprints" - structured collaborations with large enterprises that get a working AI solution deployed in days, not the industry-standard months. Customers span financial services, insurance, real estate, telecom, industrial, and compliance-heavy sectors.
