Prentis Wants AI to Do Your Office Work — Not Write Code
Ritankar Das thinks coding is overhyped. The 31-year-old founder of Prentis, an AI research lab co-founded by Reid Hoffman and Mark Pincus, has a different bet: automating routine computer tasks will soon become AI's biggest use case. And he's raising $100 million at a $1 billion valuation to prove it.
Prentis launched in April with a simple premise. Office workers spend hours every day clicking through the same screens, filling out forms, hunting down paperwork. What if an AI agent could just... do that?
The answer, according to Prentis, is Hive-32B, a model small enough to run cheaply but smart enough to navigate real Windows applications end-to-end. The company claims it handles insurance claims, customs duty refund exceptions, and other mundane office tasks without human intervention.
It's a crowded space. Anthropic, OpenAI, and Mira Murati's Thinking Machines Lab are all working on computer-use agents. Anthropic even acquired Seattle-based Vercept earlier this year, folding in its founders and shutting down the product. But Prentis says it has an edge.
The Numbers Prentis Is Throwing Around
Prentis has already signed contracts worth up to $50 million with several customers, including a healthcare management services organization, a manufacturer, and goods and clothing manufacturers, according to two people familiar with the discussions.
The pitch deck paints a more aggressive picture: an estimated $75 million annualized run rate by Q3 2026. But there's a catch. Those figures reflect "estimated annualized value based on a contracted fee equal to 20% of savings realized, not recognized revenue, and are performance-dependent and subject to final execution," according to TechCrunch's reporting.
Translation: Prentis gets paid only if it actually saves its customers money. That's a bold assumption when you're asking businesses to hand over their workflows to a system that hasn't been independently verified.
The company claims its cost per task is roughly 10 times lower than frontier API providers like OpenAI or Anthropic. If that's true, it makes economic sense to deploy these agents across everyday office workflows rather than reserving them for high-value tasks.
Benchmarks That Raise Eyebrows
Prentis says its Hive-32B model outperforms OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6 on two computer-use benchmarks: WindowsAgentArena, which measures end-to-end task completion on real Windows applications, and ScreenSpot-v2, which tests a model's ability to locate the right on-screen control.
TechCrunch hasn't independently verified these results. That matters. Benchmarks can be gamed, especially when you're the one writing the test suite. The real question is whether these models hold up in the messy, unstructured world of actual office software.
The pitch deck argues that running a much smaller, cheaper model is the secret sauce. Bigger models aren't necessarily better when you're automating routine tasks. You need speed, reliability, and low cost. Prentis claims it has all three.
The Backers Behind the Lab
Das isn't working alone. His co-founders bring serious credibility to Prentis.
Reid Hoffman, the LinkedIn co-founder and Greylock partner, is stepping down from Microsoft's board after nearly a decade to go "founder mode" on Manas AI, an AI drug-discovery startup he's also backing. He was an early OpenAI investor and co-founded Inflection AI with Mustafa Suleyman before Microsoft absorbed most of that team in 2024.
Mark Pincus, the Zynga founder, runs Reinvent Capital with Hoffman as a senior adviser. He published a memoir, "Life at the Speed of Play," last month. Both are serial entrepreneurs who've seen AI hype cycles come and go.
Das himself has an impressive background. He was UC Berkeley's youngest University Medalist in more than a century, graduating at 18 with a double major in bioengineering and chemical biology before earning a master's in biomedical engineering at Oxford. He founded Titan, a holding company that builds and operates AI companies, in 2014 after dropping out of a Cambridge AI PhD program.
Titan has funded ventures like Tala Health, which raised $100 million in seed funding last year, and Forta Health, an autism care startup that raised $55 million led by Insight Partners in 2024. Dascena, a disease prediction company, was acquired by CirrusDx in 2022.
Prentis has hired more than 25 employees, including researchers from OpenAI, Google DeepMind, Meta, Tencent, and Alibaba, according to its website.
Why Computer Use Over Coding?
Here's the core bet: automating everyday office tasks will soon outpace coding as AI's biggest use case.
Das argues that while coding has gotten all the attention, the real value lies in automating the mundane tasks that keep office workers busy. Insurance claims processing. Customs duty refund exceptions. Inventory management. These are repetitive, rule-based tasks that don't require creative problem-solving. They just need someone to click through screens and fill out forms.
The market for this is enormous. Every company has back-office workflows that could be automated. The question is whether AI agents are ready for prime time.
Prentis says yes. The company argues its models handle these tasks well enough to justify deploying them at scale. Whether that's true remains to be seen.
The Competition Is Fierce
Anthropic, OpenAI, and Thinking Machines Lab are all working on computer-use agents. Anthropic even acquired Vercept, a Seattle-based startup focused on this category, earlier this year.
Prentis has a head start in some ways. It launched in April, while competitors are still experimenting. But the market is crowded, and differentiation is hard.
Anthropic has been acquiring talent in the category directly. OpenAI has GPT-5.4, which Prentis claims its model outperforms. Thinking Machines Lab has Mira Murati leading the charge.
Prentis hasn't responded to TechCrunch's request for comment. That's unusual for a startup in fundraising talks. It might suggest caution, or it might just mean the team is focused on closing deals.
What This Means for the AI Industry
If Prentis is right, we're looking at a fundamental shift in how AI gets deployed. Instead of coding assistants and creative tools, the biggest use case becomes mundane office automation.
That's not necessarily bad. Automating routine tasks frees up human workers to focus on higher-value work. It could reduce costs for businesses and improve efficiency across industries.
But it also raises questions about job displacement, security, and the reliability of autonomous agents handling sensitive business data. These are issues Prentis hasn't addressed publicly.
The company's pitch deck suggests the market is ready. Whether the technology is ready remains to be seen. Prentis has contracts worth $50 million, but those are performance-based. If the models don't deliver, the revenue evaporates.
The next 12 months will be telling. If Prentis hits its $75 million annualized run rate target, it'll prove that computer-use agents have real commercial viability. If it misses, the hype cycle might cool down faster than expected.
One thing's clear: the race to automate office work is on. Prentis is betting big that it can win.
