San Francisco startup Freehand closed a $75 million Series B on July 29 to expand its autonomous AI agents, which already handle supply-chain spending decisions for Meta, Unilever, Pfizer, and Johnson & Johnson.
The round was co-led by Battery Ventures and NewRoad Capital Partners, with Nexus Venture Partners and former U.S. Commerce Secretary Penny Pritzker also joining. It brings Freehand's total raised to $100 million, a sharp jump from the $25 million Series A the company closed in March 2024. That trajectory alone says something about how quickly the appetite for real, decision-making AI in enterprise back offices has shifted.
What Freehand actually does is worth spelling out, because it's a fair bit more than another procurement dashboard. When a large enterprise receives an invoice, a human team historically has to cross-reference the invoice against supplier contracts, email threads, rate negotiations, and exceptions - then decide whether to pay in full, dispute the charge, or push back on price. Freehand's agents do that entire sequence autonomously. They read the contracts, parse the communications, build context across a spend category, and take action directly inside enterprise systems. The agent decides: pay, dispute, or negotiate.
The underlying technology is what Freehand calls a Category Context Graph - a proprietary structure that captures every prior decision, transaction, and exception within a given spend category. That's what lets the agents reason across messy, unstructured data, the sort that usually requires an experienced procurement manager who has been working a supplier relationship for years. The company claims clients see workflows accelerate sevenfold and spending cut by as much as 10%.
Dharmesh Thakker, General Partner at Battery Ventures, described Freehand's agents as fundamentally different from the co-pilot tools that dominated enterprise AI discourse a year or two ago. His argument, stated plainly: co-pilots assist humans; Freehand's agents possess the enterprise context, make the call, and execute it. That's not a subtle distinction. It separates a productivity tool from a labour replacement, and the investor community has clearly started pricing the difference. Gregoire Lehmann at NewRoad Capital made a similar point, noting that Freehand competes on measurable business outcomes rather than feature counts.
CEO and co-founder Nitin Jayakrishnan and his co-founder Abhijeet Manohar are not first-timers in this space. The two previously built Pando, an enterprise SaaS platform for transportation management and procure-to-pay logistics, before moving to board roles at Pando in early 2024 and founding Freehand as an agent-native entity from the start. Pando was sold to a strategic buyer in early 2026. That history matters: Freehand wasn't built by people who discovered supply chain last year. The domain expertise baked into the Category Context Graph comes from a long time working inside the messy operational reality of large-scale enterprise procurement.
The client roster makes the funding less surprising than it might otherwise look. Meta, Unilever, Johnson & Johnson, Pfizer, Cardinal Health, Dunkin' - these aren't pilot programmes or proofs of concept. They're production deployments at companies with procurement operations of a scale where a 10% reduction in supply-chain spend is not a rounding error. Forbes reported on the round the same day it was announced, noting the particular significance of Freehand emerging from stealth with this calibre of enterprise customer already live.
What the $75 million is actually for #
Freehand says the new capital will go toward scaling its AI teams and expanding the categories of supply-chain spend the platform can manage. Right now the agents focus primarily on invoice and payment operations, supplier management, and contract enforcement. More spend categories means more surface area inside each client's operations - and more agents doing work that currently sits on someone's desk.
That's the real story here. Agentic AI has been a buzzword for two years. Freehand is one of the clearest examples of it operating at production scale, inside the supply chains of some of the world's most recognisable companies, making financial decisions with real dollar consequences. The $75 million tells you investors think that's just the beginning of the market.
Also read: K2 Space is reportedly valued at $6.8 billion as Iconiq Capital bets on satellite hardware • Dili raises $21.7 million to automate the compliance paperwork slowing AI infrastructure buildout • Array Labs raises $21M from Mitsubishi Electric to build satellite radar tracking in the Pacific