BeyondMinds today announced a $15 million funding round led by Check Point cofounder Marius Nacht. The company, which is based in Tel Aviv and New York, says it will devote the bulk of these funds to sales and product R&D as it expands its customer base.
The benefits of AI and machine learning can feel intangible at times, but surveys show this hasn’t kept enterprises from adopting the technology in droves. Business use of AI grew a whopping 270% over the past four years, according to Gartner. And Deloitte says 62% of respondents to its corporate October 2018 report deployed some form of AI, up from 53% the year before. But adoption doesn’t always meet with success, as the roughly 25% of companies that have seen half their AI projects fail will tell you.
BeyondMinds was cofounded in 2018 by Or Kremer, Roey Mechrez, and Rotem Alaluf, previously a deep learning and natural learning processor researcher at Technion-Israel. The startup offers a modular AI technology stack to facilitate enterprise product deployments. Its out-of-the-box datasets and representations can be used to train computer vision, natural language, text-to-speech, and time-series models small enough to run in constrained computational environments, with modules that explain the models’ decisions.
“I previously worked at one of Israel’s leading defense companies dealing with a lot with noisy, messy, and dynamic data. Making AI systems work in these scenarios and providing accurate and stable predictions were the most time-consuming activities in building solutions,” CEO Alaluf told VentureBeat via email. “After working on many different issues, I noticed that the same problems repeated over and over. I understood early on that if we were able to generalize core production obstacles, we’d solve one of the main barriers to mass AI adoption. And that’s exactly what BeyondMinds has done.”
BeyondMinds operates like a consultancy, beginning with an AI potential assessment of business needs, data, and constraints. The next step entails defining product functionality and integrating data ahead of building and deploying the customized AI product. BeyondMinds then continuously monitors and improves the product with live data, taking into account feedback to optimize features in changing environments.
This is reminiscent of competitors like DataRobot, which recently raked in $270 million to further develop its end-to-end enterprise AI platform. But BeyondMinds claims to “stabilize” machine learning solutions in production where the data is “dynamic and noisy” by deploying trust, monitoring, and feedback technologies to achieve increasing value over time.
“We use cutting-edge technologies that enable us to create an accurate and stable solution under extreme and dynamic scenarios that contain the required ingredients for successful production,” Alaluf explained. “We have developed and implemented new technologies in the fields of explainability, confidence, bias mitigation, smart annotation and feedback, monitoring, technologies to handle small amounts of data, and methods to stabilize AI systems in production to ensure increased value after deployment. We use techniques of self-supervised learning, among others, to enable us to reduce the need for large amounts of labeled data.”
BeyondMinds serves businesses across finance, industry, and government verticals and claims to have customers in teams at Samsung, KPMG, and Microsoft, along with “numerous leading defense companies and financial institutions.” Typical applications include defect detection, predictive maintenance, fraud detection, insurance claim automation, risk assessment, and underwriting.
Grove Ventures co-led BeyondMinds’ raise, which is the startup’s first public round. BeyondMinds has raised $16 million to date and has over 60 employees.