Solid Launches with $20M to Automate AI Reliability
NYC-based startup Solid launched with a $20 million seed round to build an enterprise AI reliability platform. The company aims to automate "semantic engineering" to ensure LLM outputs meet accuracy, compliance, and reliability standards. Solid is hiring engineers to build monitoring and validation pipelines for regulated industries like finance and insurance.
- The company's co-founders, CEO Yoni Leitersdorf and CTO Tal Segalov, are both second-time founders who previously served together in Unit 8200, the Israeli Defense Forces' elite signals intelligence unit. Leitersdorf's prior company, Indeni, was acquired by BlueCat. - The $20 million seed round was co-led by the venture creation firm Team8 and the VC firm SignalFire. According to the CEO, the round's size was increased at the investors' request, signaling strong conviction in the company's approach to a significant market problem. - Solid is positioning a new discipline it calls "semantic engineering," predicting the rise of "semantic engineers" who will be responsible for systematically defining and validating business logic for AI consumption. The company suggests that today's data analysts are well-positioned to evolve into this future role. - The platform creates a continuously updated "context graph" that acts as a single source of truth for business meaning, aiming to solve the issue of different teams defining key metrics like revenue or customer activity in conflicting ways. - Solid claims its platform can increase the accuracy of AI responses from a baseline of 20-30% to over 85%, while reducing the manual effort of maintaining business definitions by 50-70%. - The company has offices in both the United States and Israel and plans to use the funding to expand its U.S. sales team and grow its R&D team in Israel. - SurveyMonkey is an early customer, with its VP of Data stating that Solid provides a foundation where business definitions stay aligned as data evolves, allowing them to trust the outputs of their AI workflows.
