Plurai secures $13M seed round to boost AI agents with simulation tech
The startup aims to raise AI agent reliability by creating digital twins and synthetic data, borrowing methods that got self‑driving cars on the road.

Plurai announced a $13 million seed investment to develop simulation‑based training for AI agents, a move designed to lift their performance to levels seen in autonomous‑vehicle development, TechCrunch reported.
How does Plurai intend to make AI agents more reliable?
Plurai builds a digital replica of a company’s electronic systems—including knowledge bases and tools—to map every possible agent workflow and generate synthetic edge‑case data for model training, according to TechCrunch. The platform then uses this simulated environment to evaluate and correct agents, working alongside existing tools such as Claude Code, Codex, Copilot and Cursor.
What funding did the startup receive and who led the round?
The seed round raised $13 million, led by venture firm Team8 with participation from U&I, Velocity and Mercer, TechCrunch said. The capital will fund the creation of the simulation layer and the post‑training of small, open‑weight models tailored to individual customers.
Who founded Plurai and what is their background?
Co‑founders Ilan Kadar and Elad Levi previously worked at Nexar, the autonomous‑driving technology company, and Kadar has experience simulating Waymo’s end‑to‑end driving software, TechCrunch noted. Kadar told TechCrunch that the surge of AI agents mirrors the early demo‑heavy phase of self‑driving tech, prompting the need for a reliable, cost‑effective simulation approach.
Additional context from the source notes that benchmark tests like LiveBench 2.0 gave Claude Fable 5.2 an 82.9 % success rate and ChatGPT‑6 Astra 75.6 % on 100 tasks, while EON’s corporate‑environment test saw Claude Opus 4.8 achieve 76.8 % accuracy. Plurai’s own trials with Nvidia and Nebius reportedly outperformed mid‑tier frontier models at lower cost, and angel investor Reah Miyara of Google DeepMind highlighted the startup’s edge in generating synthetic data to tune models for specific business policies.
This report is based on original reporting by TechCrunch. Read the original source →
Related coverage
- OpenAI rolls out ChatGPT Wallet for AI‑driven purchases to U.S. Pro users13 Nov 2026, 17:03 UTC