Reid Hoffman and Mark Pincus Back Prentis, an AI Lab Targeting $100M Raise to Automate Office Workflows

A new AI research lab called Prentis is in talks to secure $100 million in funding at a $1 billion valuation, according to two people familiar with the discussions who spoke to TechCrunch. The startup, which launched in April, was co-founded by serial entrepreneur Ritankar Das alongside tech heavyweights Reid Hoffman and Mark Pincus. Hoffman brings his experience as LinkedIn's co-founder and a prolific AI investor, while Pincus is known for founding Zynga and his early bets on social platforms. Das, the CEO, contributes deep technical leadership to the venture.

The $1 billion valuation target signals major investor confidence in a company that's barely a few months old. Prentis has already signed contracts worth up to $50 million with several customers, including a healthcare management services provider, demonstrating early commercial traction that likely underpins the aggressive fundraising terms.

Prentis is training what it calls computer use models. These are AI systems designed to observe and learn how office workers navigate routine workflows across documents and internal systems, then replicate those actions. The goal is to build AI agents that can control computers directly, clicking through interfaces, filling forms, and moving data between applications without human intervention.

The startup’s thesis is that automating routine computer tasks will soon outpace coding as AI's biggest use case. Instead of writing code to automate a process, Prentis agents watch a human do the task once and then take over. This approach targets the sprawling, unglamorous work that fills white-collar days: processing insurance claims, reconciling spreadsheets, and chasing down paperwork for customs duty refund exceptions.

Computer use agents represent a shift from AI that generates content to AI that performs actions. Rather than outputting text or code, these models interact with graphical user interfaces the way a person would, clicking buttons, filling fields, and navigating menus. This capability doesn't require APIs or backend access, it works on top of existing software.

For enterprises, the appeal is straightforward: millions of employees spend hours each week on repetitive computer tasks that don't require judgment, only persistence. Prentis is betting that automating this layer of work delivers faster ROI than building custom integrations or retraining staff. The company's early contracts in healthcare and logistics suggest demand is real, not theoretical.

Reid Hoffman’s involvement gives Prentis a direct line to Silicon Valley’s power networks. As LinkedIn’s co-founder and a partner at Greylock, Hoffman has backed some of the most consequential AI companies of the past decade. Mark Pincus, the Zynga founder, brings product intuition and a track record of building consumer-scale platforms. Ritankar Das rounds out the leadership with technical and operational depth.

This isn’t Hoffman’s first AI venture. He previously co-founded Inflection AI with Mustafa Suleyman and Karén Simonyan, a company that raised over $1 billion before pivoting to enterprise AI. Prentis appears to be a more focused bet, zeroing in on a single paradigm rather than building a general-purpose assistant.

Prentis enters a market that’s already crowded. Companies like Adept, backed by $350 million in funding, are building similar computer use agents. OpenAI and Anthropic have both signaled interest in models that can control software. Google’s DeepMind has published research on GUI-navigating agents. Prentis differentiates by targeting specific enterprise workflows from day one and signing revenue contracts before closing its first major funding round.

The lab’s approach also sidesteps the developer-tooling market entirely. Instead of selling to engineering teams, Prentis sells to operations leaders who want to automate without writing code. If the thesis holds, computer use models could eat into the RPA market and reshape how enterprises think about automation budgets.

Prentis is still in talks for the $100 million raise, meaning the deal isn’t finalized and terms could shift. The company’s early customer contracts provide a revenue base that most AI labs at this stage lack, which should strengthen its negotiating position. If the round closes at the reported valuation, Prentis will be one of the fastest AI startups to reach unicorn status from launch.

The broader signal is worth watching. When a founder of Hoffman’s caliber commits to computer use models over generative coding tools, it suggests a conviction that the next wave of AI value won’t come from helping developers write code, but from replacing the work that happens after the code is already written. Enterprise buyers and investors will be watching closely.