Plato Raises $16 Million to Build Simulated Workplaces Where AI Agents Practice Real Jobs
The San Francisco startup pays professionals to recreate their jobs as resettable simulations that AI labs use to train and test agents.
Published
Last updated / Fact-checked on
Plato, a San Francisco startup that builds simulated workplaces for AI agents, has raised a $16 million Series A. Nava Ventures led the round, announced Oct. 28. Gradient Ventures, Transpose Platform and Microsoft AI CEO Mustafa Suleyman also took part. Nava founder Freddie Martignetti joins Plato's board. The round brings Plato's total funding to $19 million, following a seed round led by Gradient Ventures.
What Plato builds
Founded in 2025, Plato builds working copies of the software, tasks and situations that make up real jobs. It currently focuses on customer support, people operations, finance, accounting and healthcare administration. AI labs use these environments in two ways. One is reinforcement learning, where a model tries a task, gets feedback and improves. The other is evaluation.
The point is safety and repeatability. As Plato's release explains, a model learning to place an order in live software could spend real money. One practicing an admin task could change real records. Plato's simulations can be reset and reused, so labs can let models fail without consequences.
"Mastering real complex work comes from experience doing those tasks," said co-founder and CEO Robert Farlow. "We're building simulations that capture the tools, decisions, and unexpected situations those tasks involve, so AI models can gain experience in safe, contained environments before developers put them to work."
Grading the process, not just the score
Plato says it grades both the result and how a model got there. Completing a purchase doesn't count as success, for example, if the model went over budget. Its evaluations also look for models that exploit weaknesses in a simulation to earn a high score.
"The feedback a model receives shapes what it learns," said co-founder and CTO Pranav Putta. "We build evaluations that reflect what the job actually requires, so developers can distinguish a high score from work done correctly."
Where the money goes
Plato will use the funding to pay professionals to recreate their own jobs as simulations. It will also deploy AI agents on real tasks to gather more experience.
| Detail | Information |
|---|---|
| Round | $16M Series A |
| Lead | Nava Ventures (Freddie Martignetti joins board) |
| Other investors | Gradient Ventures, Transpose Platform, Mustafa Suleyman |
| Total raised | $19M |
| Founded | 2025, San Francisco |
| Focus | Support, people ops, finance, accounting, healthcare admin |
Why it matters
AI is moving from answering questions to taking actions inside business software. That raises the bar for reliability. A strong score on a static benchmark doesn't show whether an agent will respect a budget, follow a policy or recover from an unexpected situation.
Plato isn't alone in this space. The Next Web notes that AI training firm micro1 said this month it will spend $1 billion on company data to build similar practice environments. In June, Patronus AI raised $50 million to stress-test agents in simulated settings. Suleyman's involvement is also worth noting. TNW reports he has argued that AI should be provably under control before it is released.
Our analysis
Plato's bet is that environments built with working professionals capture things synthetic data misses, like edge cases, unwritten rules and judgment calls. Grading the process as well as the result addresses a known failure: agents that look successful while breaking the rules.
Some questions remain open. A simulation is only as good as the experts who describe the job and the fidelity of the copied tools. Building many detailed job environments is slow and expensive. Labs will still need to measure whether success in a simulation carries over to live systems.
For companies planning to put agents into support, finance or HR, the takeaway is practical. Ask how an agent was tested, not just how it scored. Look for evidence that it was graded on following the rules and limits of the job.
How we reported this
We checked funding details, investors and quotes against Plato's press release, and cross-checked them with The Next Web, Axios Pro and FinSMEs. Plato did not disclose customers or performance metrics, and we don't report any.
Sources
- Plato Technologies via PR Newswire, "Plato Raises $16 Million to Build the Training Grounds for AI's Next Leap Into Real Work" (Oct. 28, 2026): https://www.prnewswire.com/news-releases/plato-raises-16-million-to-build-the-training-grounds-for-ais-next-leap-into-real-work-302921832.html
- The Next Web, "Plato raises $16M to build fake offices where AI agents learn real jobs" (Oct. 28, 2026): https://thenextweb.com/news/plato-16m-series-a-simulated-workplaces-ai-agents
- Axios Pro, "Workplace simulation startup Plato raises $16M" (Oct. 28, 2026): https://www.axios.com/pro/enterprise-software-deals/2026/10/28/workplace-simulation-plato-16-million-nava
- FinSMEs, "Plato Raises $16M in Series A Funding" (Oct. 2026): https://www.finsmes.com/2026/10/plato-raises-16m-in-series-a-funding.html
DailyMech used AI tools to help research and draft this article; it was fact-checked by our editors.