AI startup founders say real value, not the AI label, must drive the pitch
At an AI Week panel hosted by Calcalist and Google, startup founders and an industry executive argued that AI companies must solve real problems, not rely on the technology buzzword. The discussion was moderated by Elihai Vidal of Calcalist and featured Rafael Broshi, cofounder and CEO of Notch; Stav Levi Neumark, cofounder and CEO of Alta; and Sefi Tzarfati, head of partnerships, startups and funds at Google Israel.
Broshi said, “If your pitch does not work when you take out the word AI, your product is not good.” He described Notch as a platform for building agents for operations and customer service in highly regulated sectors, especially financial institutions and telecoms. Notch began in insurance, where the team first built a non-AI product and later, after ChatGPT 3.5 launched in November 2022, created an internal customer service tool for U.S. policyholders. After a few months, the company shut down the insurance product, turned the tool into a product, and sold it to other companies.
Levi Neumark said Alta was founded shortly after ChatGPT 3 was released, as new models and agents emerged. Alta builds agents for go-to-market teams, using a “company brain” and agents for BDRs, sales staff and customer success managers. She said companies must decide where to draw the line between AI and human work based on risk versus return, and that Alta helps customers identify safer, higher-value use cases. “Companies that come with willingness also see immediate improvements in performance,” she said.
Tzarfati said the main challenge is building trust in models and products. Google for Startups helps connect startups with large companies and adapt products to enterprise needs. He said the new rule is specialization, because focused startups are harder to replace and easier for customers to adopt. Broshi said the main moat in the AI era comes from regulation or accumulated domain knowledge, but the biggest challenge is the “last mile,” connecting fast-moving technology to a company’s actual business and organizational structure.
The panel concluded that startups should build around a problem, stay agile as models change every month or two, and remain close to technology and customers. Levi Neumark said this is “the best and worst time to start a startup,” while Tzarfati called it a historic wave with room for many niche players. Broshi ended with the advice to solve a problem you understand deeply.
