Convey lets non-technical teams build AI “teammates” by walking through processes on screen and turning them into versioned, testable programs that run reliably. a16z led Convey’s $38M Series A after its agents logged over 1.1 million work hours at NBCUniversal, TelevisaUnivision and others, freeing up hundreds of hours weekly on reporting and ad ops.

• Convey turns non-technical employees' screen-recorded workflows into versioned, testable AI agents rather than ad hoc prompts, making them reliable enough for business-critical processes.
• Its agents have logged over 1.1 million work hours at companies like NBCUniversal, Unity, and TelevisaUnivision, with one streaming service saving 450+ hours weekly and Savoya boosting EBITDA 40% YoY.
• a16z led Convey's $38M Series A based on these results and the founders' prior track record (including automating a critical manual matching role at DoorDash).

Enterprises struggle to test AI forecasts in real-world conditions, so startups are using prediction markets as a live sanity check. Augur lets companies spin up private markets where employees trade on AI-generated predictions to catch model flaws before they cause costly errors. It monetizes through tiered SaaS plans, transaction fees on public markets, and a data API for aggregated market sentiment.

• Augur runs private, real-money/token prediction markets where employees bet against a company's own AI forecasts to expose model blind spots that backtesting misses.
• Revenue comes from tiered SaaS pricing on private markets, transaction fees on public markets, and eventually an API selling anonymized sentiment data to hedge funds.
• Growth tactics include a public demo tied to high-profile events (like Fed decisions) for SEO, an open-source engine on GitHub, and a "Forecast Grader" tool to hook clients.
• The whole platform is buildable fast with a lean stack (Node.js, Supabase Realtime/Socket.io, PostgreSQL, Next.js/Tailwind), letting a small team ship it in weeks.