FIG. 01 — AGENTS AT WORK, 2026
Rise2 is a San Francisco lab working on the two things autonomy actually runs on: agentic systems that do real work, and the training data that teaches machines to move. Plus the education that makes teams fluent in both.
Autonomous intelligence will do the work — in software and in the world. We build it, deploy it, and teach organizations to wield it.
Agents that reason, plan, and execute are moving from demos to production. Robots that learn are moving from labs to factory floors. Neither is blocked by model architecture anymore — they're blocked by data, evaluation, and the people who know how to wield them.
So that's where we work: agentic systems built and hardened for enterprises, training data and benchmarks for embodied AI, and intensive programs that turn product teams into builders. We publish the research openly as it matures.
Autonomous agents that reason, plan, and execute — designed, built, and deployed into enterprise workflows where they do real work.
Demonstration capture, curation, and simulation — the episodes that teach manipulation policies what contact, failure, and recovery actually look like.
Held-out task suites and scoring rubrics that tell you whether a policy is ready for a real floor, before it gets there.
Intensive, hands-on programs that turn product teams and executives into AI-fluent builders who ship, not just learn.
We're entering the age of autonomous AI systems — agents that don't just answer, but act. We design, build, and deploy them for enterprises.
Agents that reason, plan, and execute complex tasks independently.
Connected to your existing tools, APIs, and databases from day one.
Teams of specialized agents orchestrated toward complex objectives.
Conversational experiences that understand context and intent.
Systems that improve through feedback, evaluation, and optimization.
Define the objective; the agent finds the optimal path.
Upskill yourself or your entire organization — from solo practitioner tracks to team-wide, enterprise programs, all built around shipping real AI products.
The PM and founder's playbook to ship working products with AI-first development. Build and deploy a real product by the end.
View program →Executive-level fluency in autonomous AI systems — how they work, where they pay off, and how to lead their implementation.
View program →Tailored programs for startups, scale-ups, and enterprises that turn AI ambition into shipped products across the org.
View program →Papers, benchmarks, agent frameworks, and teaching material from our work in agentic and embodied AI — released as they mature. Progress this consequential shouldn't advance in the dark.
Get notified on releases →