What Freehand sells
Freehand, a San Francisco-based startup, builds autonomous artificial intelligence (AI) agents that take over back-office supply-chain work: auditing invoices and deciding whether to pay them, negotiating with suppliers, running procurement, and reconciling data across the patchwork of enterprise systems a large company typically runs. On July 29, the company announced $75 million in new funding, alongside a customer list that includes Meta, Unilever, Johnson & Johnson, Pfizer, Dunkin', and Cardinal Health.
The company was co-founded by chief executive Nitin Jayakrishnan, who previously built and sold the enterprise logistics software company Pando, together with Abhijeet Manohar. Crunchbase News dates the company's founding to February 2024, while Forbes describes it as having operated in stealth since 2023 before emerging publicly this year. Battery Ventures and NewRoad Capital Partners co-led the round, with Battery partner Dharmesh Thakker joining Freehand's board. Also participating were PSP Growth, the investment fund of former US Commerce Secretary Penny Pritzker, and Nexus Venture Partners.
A funding round two outlets can't agree on
Here is where the story gets less tidy than the press release suggests. Crunchbase News reports this as a Series B round, bringing Freehand's total funding to $100 million. Forbes and Retail Technology Innovation Hub both describe the same $75 million raise as a seed round. Neither framing is presented as tentative; both are stated as fact by outlets that normally get funding-stage labels right; and Freehand's own press materials, so far as they were reviewed for this piece, do not resolve the discrepancy.
The gap matters beyond terminology. A $75 million seed round for a company with named Fortune 500 customers and multiple institutional co-leads would be unusually large even by 2026 standards, and would suggest a company positioning itself as earlier-stage than its customer list implies. A $100 million cumulative total across a seed and a Series B is a more conventional trajectory for a two-year-old enterprise startup with this level of traction. Readers should treat the round's official stage, and Freehand's true funding history, as unconfirmed rather than take either outlet's label at face value.
The pitch, in the company's own words
Freehand frames its market opportunity around a specific piece of arithmetic from its own materials: Jayakrishnan says enterprises spend $16 billion a year on supply-chain software and $348 billion a year hiring people to do the work that software cannot, and Freehand's agents are built to close that gap. Forbes reports the company's pitch that its agents handle 99% of decisions without human intervention, targeting what it describes as a $20 trillion US supply-chain market. Jayakrishnan put the company's positioning bluntly: "Our AI agents will decide whether to pay an invoice, act on that decision and take accountability."
On customer results, Freehand says clients recover 5 to 10 percent of spend in complex purchasing categories, complete workflows five to seven times faster, and cut procure-to-pay cycle times by more than 70%. Those figures come from the company and its investors rather than from independently audited case studies, and none of the outlets covering the raise appear to have verified them directly. Unilever's global vice president of supply chain, Matt Algar, is quoted describing the deployment as "one of the first full-scale agentic deployments at Unilever," which is a more measured claim than the percentage figures and is at least directly attributed to a named customer executive.
Freehand is one of a growing number of startups pitching AI agents as a replacement for, rather than an addition to, the back-office headcount that has traditionally run procurement and accounts payable. That positioning is part of a wider shift enterprise software buyers are already navigating this year, as vendors across categories reframe per-seat software pricing around autonomous agents doing the work a team used to do. What separates a credible entrant in that field from a marketing exercise is usually the willingness of paying customers to describe results in specific, checkable terms rather than round percentages supplied by the vendor, which is exactly the gap between Algar's quote and the 5-to-10-percent and 70-percent figures in Freehand's own materials.
What it means in practice
The raise is real, the investor names are credible, and the customer list is large enough that this is not a startup selling vaporware to nobody. But the seed-versus-Series-B discrepancy is a useful reminder that funding announcements are marketing events first and factual disclosures second, and that outlets repeating a company's own figures, whether about round size or about how much spend its software recovers, are not the same as those figures being independently verified. Any procurement or finance team evaluating Freehand, or a competitor making similar claims, should ask for reference customers who will speak on the record about actual results, not just the percentages in a press release, and should treat "handles 99% of decisions" as a claim to test against their own data rather than a settled fact.
