WithCoverage, an AI-enabled insurance startup backed by Sequoia and Khosla, has built its entire engineering workflow around autonomous AI coding agents.
JD Ross, co-founder of real estate iBuyer platform Opendoor, just said something that should make every software engineer mid-keystroke. At his new startup WithCoverage, not a single full-time engineer writes code by hand. Every line ships through AI tools like OpenAI’s Codex, Anthropic’s Claude Code, or Cursor IDE.
Ross made the disclosure on Joe Lonsdale’s American Optimist podcast, and the claim is striking not because AI-assisted coding is new, but because this is a venture-backed company with over $43 million in funding treating manual code writing as a legacy practice. In English: the engineers don’t type code, they manage fleets of AI agents that do it for them.
How the workflow actually works #
WithCoverage, co-founded by Ross and Max Brenner, operates in the AI-enabled risk management and insurance space. The engineering team’s job has essentially been redefined. Rather than sitting in an IDE hammering out functions and debugging syntax errors, engineers manage multiple autonomous agents that handle the full pipeline from ticket to pull request.
The result, according to Ross, is that feature delivery timelines have compressed from months to days. Ross also suggested that the cost of software development is dropping so dramatically it could approach zero.
WithCoverage has secured its Series B round led by Sequoia Capital and Khosla Ventures, pushing total funding past $43 million.
Why this matters beyond insurance #
WithCoverage is an insurance startup, not an AI company. It pairs expert teams in risk and insurance with advanced AI tooling to serve enterprise customers. We’ve already seen crypto-native teams experiment with AI coding tools. But WithCoverage’s approach goes further. It’s not AI-assisted. It’s AI-exclusive. Engineers don’t touch the code. They review, direct, and deploy what the agents produce.
What this means for investors #
The risk, of course, is quality and security. AI-generated code isn’t immune to bugs, and in industries like insurance and finance, where a single vulnerability can be catastrophic, the review and auditing layers become even more critical. Ross’s model accounts for this by keeping skilled engineers in the loop as supervisors.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our