Freehand raises $75m to deploy AI agents across enterprise supply chains
Freehand, a San Francisco-based startup building autonomous AI agents for supply-chain spend management, has closed a $75 million funding round co-led by Battery Ventures and NewRoad Capital Partners. The raise also drew participation from PSP Growth, the venture vehicle of former US Commerce Secretary Penny Pritzker, alongside Nexus Venture Partners and undisclosed others. Battery Ventures general partner Dharmesh Thakker is joining Freehand's board as part of the deal.
The company emerged from stealth alongside the announcement, disclosing active deployments at Meta, Unilever, Johnson & Johnson, Pfizer, Dunkin' and Cardinal Health. Across those early customers, Freehand says clients recovered between 5 and 10 per cent of spend in targeted categories, completed procurement workflows five to seven times faster, and cut procure-to-pay cycle times by more than 70 per cent. The company did not name an independent auditor for those figures, and they should be treated as vendor-reported benchmarks.
What Freehand actually does
At its core, Freehand replaces the outsourced business-process teams and legacy procurement software that handle invoice auditing, supplier negotiations, contract enforcement and payment reconciliation. Its proprietary Category Context Graph ingests structured data from enterprise systems alongside unstructured content in contracts, emails and documents, building a knowledge base the company says gives its agents the situational awareness of an experienced supply-chain practitioner. Each agent interaction feeds back into the graph, compounding the system's accuracy over time.
Co-founder and chief executive Nitin Jayakrishnan, who previously built and sold Pando, a logistics SaaS platform, framed the market opportunity with two statistics from the release: enterprises spend $16 billion a year on supply-chain software and a further $348 billion on the human labour required to compensate for what that software cannot do. "We built Freehand to close that gap," he said, "with AI agents that decide, act and take accountability for outcomes."
Matt Algar, global vice-president of supply chain at Unilever, offered the most substantive customer endorsement in the release, describing Freehand as "one of the first full-scale agentic deployments" at the company and characterising it as an early signal of a shift from software that assists to software that runs supply-chain operations directly.
Market context and competitive positioning
Freehand enters a procurement-automation market that already features well-funded competitors across several sub-categories: Coupa and Jaggaer dominate legacy source-to-pay; newer agentic entrants such as Zip and Pactum are attacking supplier negotiation and intake workflows; and the major ERP vendors, including SAP and Oracle, are embedding AI assistants directly into their procurement modules.
The company's vertical-wedge positioning, focusing narrowly on spend audit and payment operations rather than the full procurement suite, is consistent with a pattern seen in other enterprise AI entrants: land in one high-value, well-defined workflow, prove measurable ROI, then expand horizontally. Whether Freehand can defend that wedge as ERP vendors deepen their own agentic capabilities is the central competitive question investors and prospective customers will watch.
The funding announcement also arrives at a moment of heightened attention to supply-chain resilience. Tariff volatility and shifts in cross-border outsourcing economics are pushing large enterprises to scrutinise their third-party spend more aggressively, a macro tailwind Freehand cites directly as a driver of customer urgency. Buyers evaluating agentic procurement tools will, however, need to satisfy themselves on data-governance and audit requirements: systems that execute payment decisions autonomously inside enterprise ERP environments will face close scrutiny under internal controls frameworks and, for regulated industries, sector-specific compliance obligations. Freehand's emphasis on a fully auditable decision trail from the Category Context Graph appears designed to address precisely that concern.