The core message of the provided portion of this fireside chat is that Wonderful is positioning itself as an enterprise AI company built to deploy agents inside large global corporations, with a strategy shaped by speed, top-down selling, and a belief that AI will remove many traditional software constraints. The moderator frames the session around a question both individuals and corporations are asking: where to start with AI, how to deploy a first agent, and how to make that agent useful in daily work. Wonderful is introduced as the company addressing that enterprise version of the problem, with founder Bar Winkler described as operating at “hyper speed” after raising major capital in less than two years, while Index Ventures’ Sahir Azam is introduced as an early and continuing investor in the company.

Sahir begins by grounding the discussion in Bar’s background. Bar says he grew up in Israel and was not especially interested in technology as a child; he describes himself instead as playful and more focused on friends than school. His first founder-like experience came at age 15, when he was playing drums and convinced city hall to fund a space with equipment where people could play. He says that space is still operating in his hometown. Later, at 22, he joined IronSource as roughly employee number 20, taking on a vague, general-purpose role to help start what became the mobile division. He explains that he enjoyed ambiguity, built a small team, and saw that group become one of IronSource’s largest divisions, which deepened his interest in technology and what could be built around it.

The conversation then shifts to the path toward Wonderful. After seven years at IronSource, Bar founded approve.com, a procurement SaaS company selling to mid-market U.S. businesses. He characterizes it as “not the sexiest startup” but says he learned a great deal, raised some capital, and sold the company after two years in what he considered a successful journey. That experience clarified what he wanted next: a company that would sell top-down into large enterprises, work on complex strategic projects, and avoid being perceived as a small vendor. He explains that selling to middle management and mid-market buyers felt inefficient because the effort and energy required were high relative to the rewards compared with enterprise selling.

Bar then explains the early thesis behind Wonderful. He says the company started on January 5, less than two years before the session, with a broad view that “everything” would be affected by AI and that the team should focus on whatever allowed them to capture value in the next weeks and months. When GPT-4 real time became available in alpha to developers, Wonderful was among the first accounts to use it and saw real-time voice agents as the biggest use case. The team sold these agents to several large enterprises and secured seven-figure deals through Gmail, which Bar describes as strange enough to suggest there was a market “glitch” they could exploit. That led to deeper thesis development: if software soon has no technological limiting factors, then the logical ambition is to build an “everything platform.” The provided transcript ends at this point, so later portions of the 24:50 session are not available for analysis.