Supply chain · Deep dive

Freehand

Autonomous AI agents that run supply-chain spend for Fortune 500 enterprises — auditing freight and supplier invoices, negotiating billing disputes, enforcing contracts and processing payments inside ERP systems, replacing the outsourced back-office teams that have done this work for decades.

emerging

The question that decides it: Freehand's agents audit, dispute and pay Fortune 500 supply-chain invoices with, per the company, no human in the loop, reasoning over a Category Context Graph assembled from contracts, invoices and ERP records the customer already owns. Does the reported 5-10% spend recovery (Jul 2026) persist once the first year's accumulated billing leakage is cleaned out — recurring value rather than a one-time audit dividend — and does fully autonomous payment survive its first material error at a Meta or a Pfizer, before SAP, Oracle and Coupa bundle agentic invoice audit into the ERP and procurement systems of record Freehand's agents merely write into?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Freehand builds autonomous AI agents — it calls them AI Teams — that run supply-chain spend for very large enterprises: auditing freight and supplier invoices against contracts, disputing and negotiating billing discrepancies with vendors, coding and reconciling transactions, and processing payments inside existing ERP and procurement systems. Founded in February 2024 by the two founders of logistics-software company Pando, it operated in stealth for two years, surfaced at Manifest in February 2026 with Fortune 500 deployments already live, and announced a $75M Series B co-led by Battery Ventures and NewRoad Capital Partners on July 29, 2026 — $100M raised in total. It claims roughly 50 customers — Meta, Unilever, Johnson & Johnson, Pfizer, Dunkin’, Cardinal Health — and billions of dollars in payments processed autonomously across 60-70 countries. It matters as one of the first real tests of whether agentic AI can replace not software but outsourced labor at companies that cannot tolerate a wrong payment.

Founding story

Nitin Jayakrishnan and Abhijeet Manohar have been building the same product, in some sense, since 2016. Their first venture, iDelivery, was an Indian digital freight marketplace whose customers — Philips and ITC among them — kept asking for features that belonged inside their existing supply-chain systems. They shut it in July 2017 and founded Pando in Chennai: a no-code logistics platform that grew into a shipper-side TMS and procure-to-pay system of record, raised a $9M Chiratae-led Series A (January 2020) and a $30M Iron Pillar-led Series B (May 2023) for roughly $45M total, and signed Nestle, P&G and Johnson & Johnson.

Then came the unusual part. In early 2024, with Pando still operating, both founders stepped back to board seats and started Freehand as a separate, independent company — Jayakrishnan’s stated logic (Crunchbase News, July 2026): rather than drag an “archaic dinosaur of an industry” onto a sunsetting technology paradigm, leapfrog it onto a rising one. Pando ran under a new executive team until it was sold to an unnamed strategic buyer in early 2026, a complete shareholder exit. Read generously, this is founder-market fit distilled: eight years of watching Fortune 500 back offices check invoices, restarted agent-native with no legacy code. Read skeptically, the founders left their Series B-funded company to rebuild its most valuable module — freight audit and payment — in a new cap table, and at least one shared backer (Nexus) and one shared customer (Johnson & Johnson) moved with them. The strategic buyer and Pando’s sale price were never disclosed.

How it works

The concrete problem: a global supplier invoices a manufacturer $16.9M for six months of logistics services. Whether that number is right depends on contracted rates, surcharges, accessorials, detention clauses, SLA penalties and shipment records — historically checked by large in-house or offshore BPO teams. Jayakrishnan’s description of the job (July 2026): establish the bill should have been $16.4M, and get the supplier to concede the difference without torching the relationship.

Freehand’s agents do this end to end. The substrate is what the company calls its Category Context Graph: a semantic layer unifying contracted rates, carrier invoices, shipment records, GL rules, cost-center structures, SLA terms, payment history and past exceptions — structured ERP data plus the unstructured data buried in PDFs, emails and chats. Agents read the invoice, match line items against the contract and actual operational milestones, flag leakage, draft and conduct the dispute correspondence with the supplier, apply GL coding, accrue from actuals, post reconciled results to the ERP and release payment. Every decision carries an audit trail, and each transaction feeds back into the graph — the claimed compounding effect. The company says this runs across 60-70 countries and hundreds of currencies “without human supervision” (July 2026), a phrase that is both the pitch and the risk. Deployment is inside existing systems — ERP, procurement, collaboration tools — not a rip-and-replace.

Product and business overview

Freehand sells by spend category rather than by module. The anchor wedge is freight audit and payment — its freight product claims a closed loop from RFQ to paid invoice, recovering 3-5% of freight spend (company site, accessed July 2026) — extended to direct materials, MRO, and services/labor. Around the audit core sit sourcing and rate-negotiation agents, supplier management, contract-compliance enforcement, and reconciliation/accrual automation pitched at finance VPs. The consistent framing since the February 2026 debut is “AI Teams”: bundles of role-shaped agents deployed against a category, replacing a defined block of outsourced work, rather than a copilot assisting a human. Battery’s Dharmesh Thakker characterized the positioning as targeting the largest Fortune 500 shippers directly rather than the freight intermediaries most logistics AI startups chase (July 2026).

Business model and pricing

No pricing page exists and no rate card has surfaced; contracts are negotiated enterprise deals. The economic pitch is displacement of two budget lines Jayakrishnan sizes at $16B a year of supply-chain software and $348B of people hired to do what that software cannot (GlobeNewswire, July 2026) — i.e., Freehand prices against BPO contracts and headcount, not against SaaS seats. The company advertises outcomes — 5-10% of spend recovered in complex categories, procure-to-pay cycles cut 70%+ — which is the language of gain-share or outcome-linked pricing, but Freehand has not confirmed its actual revenue mechanics, and that opacity matters: a vendor paid on recovered leakage faces shrinking recoveries as the backlog cleans up, while a flat subscription must justify itself after year one. How revenue is booked is, for now, unverifiable.

Traction over time

Every number in that table is company-sourced. No ARR, headcount, retention or valuation figure has been published; at 29 months old with $100M raised, the gap between elite logos and zero disclosed financial evidence is what diligence would have to close.

Market analysis

The market can be sized three ways, and Freehand’s story depends on which you believe. Narrow: the freight audit and payment services market is small — Mordor Intelligence pegs it at $0.97B in 2025 growing ~14.2% annually to $1.89B by 2030, with other estimates scattered from $0.8B to $6.8B (2025), a spread that itself signals a fragmented, definitionally fuzzy category. Middle: supply-chain software spend, which Battery’s Thakker put at just over $20B a year (July 2026). Wide — and this is Freehand’s actual pitch: the $348B Jayakrishnan says enterprises spend annually on the people doing what software cannot, sitting inside the $20T+ American companies spend on supply-chain inputs per the Bureau of Economic Analysis (cited July 2026). The structural forces are real: tariffs, trade-route disruption (Red Sea, Strait of Hormuz) and immigration policy are straining the offshore outsourcing model, while agentic AI for the first time makes labor — not software — the displaceable cost. The countervailing history is sobering: supply-chain software has repeatedly attracted capital ahead of reality — Goldman-backed Slync collapsed after its CEO’s 2024 fraud conviction — a reminder that this sector’s claims deserve verification before valuation.

Competitive intel

Freehand fights on three fronts. Systems of record: Coupa (Thoma Bravo, $8B take-private, Feb 2023) and SAP Ariba (acquired 2012, $4.3B) own enterprise spend workflows and are bundling AI agents into platforms customers already pay for — the classic bundling threat to any startup whose agents write into someone else’s system. Modern procurement workflow: Zip ($2.2B valuation, Oct 2024) owns intake and orchestration at 2,000+ companies; Levelpath (~$100M raised, $55M Series B led by Battery in June 2025 — the same Battery that co-led Freehand’s B, suggesting the firm sees them as adjacent, not overlapping) rebuilds sourcing AI-native. Both approach the procure-to-pay loop from the sourcing end while Freehand enters from the invoice-audit end; collision is a matter of time. The freight wedge itself: Loop ($35M Series B co-led by J.P. Morgan Growth Equity, Oct 2023; J.P. Morgan Payments partnership, Jun 2024) automates freight audit and payment with bank distribution, and the legacy processors — Cass Information Systems, Trax — hold decades of trust and payment rails while embodying the manual cost structure Freehand attacks. Freehand’s differentiation claims: category depth (complex logistics, direct materials and MRO rather than cards and T&E — its explicit contrast with Ramp), Fortune 500 shippers over intermediaries, and full autonomy including supplier negotiation. Its exposure: every incumbent above has more distribution, and the ERPs have the data.

History and evolution

• 2016 — Jayakrishnan and Manohar launch iDelivery, a freight marketplace in India; customers include Philips and ITC.
• Jul 2017 — iDelivery closed; Pando founded in Chennai as enterprise logistics software.
• Jan 2020 — Pando raises $9M Series A led by Chiratae Ventures; clients include Nestle, P&G, Johnson & Johnson.
• May 2023 — Pando raises $30M Series B led by Iron Pillar with Uncorrelated Ventures (~$45M total); positions as AI-powered fulfillment platform.
• Feb 2024 — Founders step back to Pando board seats; Freehand incorporated in San Francisco as a separate, agent-native company.
• Mar 2024 — $25M Series A raised in stealth.
• Early 2026 — Pando sold to an unnamed strategic buyer; complete shareholder exit for founders.
• Feb 2026 — Freehand debuts publicly at Manifest with Fortune 500 deployments live.
• Jul 29, 2026 — $75M Series B co-led by Battery Ventures and NewRoad Capital Partners; Thakker joins the board.

What people say

The case for. Reference customers speak on the record, rare at this stage: Unilever’s global VP of supply chain, Matt Algar, called Freehand one of Unilever’s first full-scale agentic deployments — software that runs, not assists, the supply chain (July 2026). NewRoad’s Gregoire Lehmann cited measurable outcomes as the reason to invest; Thakker cited “tremendous customer love” and expansion within accounts (July 2026). The claimed metrics are specific: 5-10% of spend recovered, workflows 5-7x faster, procure-to-pay cycles down 70%+, AP cycle times down 80-90% in early deployments (Feb 2026). Penny Pritzker’s PSP Growth backing adds an industrial-policy imprimatur.

The complaints. There is no independent review surface at all: no G2 or Capterra listings, no Glassdoor reviews, no published customer case studies with verifiable numbers — after two years in stealth, every public data point about Freehand originates from the company or its investors, including the 50-customer and “billions processed” figures. The predecessor record offers texture: Pando’s G2 reviewers praised usable freight tracking but recurringly complained of slow performance and a dated UI (accessed July 2026), and Pando’s Glassdoor (4.2 stars, 154 reviews) includes accounts of sudden layoffs with little communication or severance. Skeptics will also note: founders exiting a venture-backed company to rebuild its roadmap next door, “no human supervision” payment claims not yet publicly stress-tested by auditors, and a spend-recovery pitch whose value is structurally front-loaded. None of it disqualifying; none of it verified.

Outlook: the open question

For Freehand to work, two things must prove durable by roughly 2028: the recovered-spend number and the autonomy claim. Concretely — the 5-10% recovery reported in July 2026 must persist after the first audit sweep clears years of accumulated billing leakage, converting a one-time dividend into recurring governance value customers will pay for indefinitely; and fully autonomous auditing, disputing and paying must survive its first material, public error at a Meta, Pfizer or Unilever without those customers reinserting humans and demoting Freehand to a copilot priced accordingly. If both hold, the Category Context Graph becomes the kind of compounding, category-specific data asset ERPs cannot easily replicate, and the BPO-displacement TAM ($348B of labor by the CEO’s own count) dwarfs any software comp — with the Fortune 500 shipper back office as defensible high ground Zip, Levelpath and Loop all approach from weaker angles.

What would sink it: the graph turns out to be a feature of the record, not a product. Freehand’s agents reason over contracts, invoices and payment history that live in SAP, Oracle and Coupa; if those vendors ship good-enough agentic audit bundled into systems customers already run — and all three are racing to — the integration-cost argument flips against the startup. The tells to watch: whether Freehand ever publishes ARR or retention rather than customer-sourced percentages; whether reference customers renew and expand into second and third spend categories in 2027; whether pricing surfaces as gain-share (revenue that shrinks as it succeeds) or subscription (which must survive the post-cleanup ROI conversation); and whether a payment error or audit dispute at a marquee account becomes public. The logos are real, the wedge is real, and the entire financial case currently rests on the company’s own numbers.

How a challenger would attack it

Weaponize verification against a company built on its own numbers. Every public data point about Freehand — 50 customers, 5-10% recovery, “billions” processed — is company- or investor-sourced, with no G2 presence, no case studies with auditable figures, and no disclosed ARR or pricing. A challenger attacks with radical transparency: published pricing, third-party-audited recovery benchmarks, SOC-style attestations on agent decisions, and named reference deployments with verifiable savings — turning Freehand’s stealth-era opacity into a procurement liability at exactly the risk-averse Fortune 500 CFOs both are selling. The second vector is the autonomy claim itself: “no human supervision” over payment release is a feature until the first material error, so a rival selling graduated autonomy — human sign-off above a threshold, full audit trails, insurance-backed error guarantees — offers the same labor displacement with a story a controller can defend to an audit committee. Third, the wedge is structurally front-loaded: 5-10% recovery comes largely from clearing accumulated leakage, so a challenger can offer the initial audit sweep free or pure gain-share, monetizing only the recurring governance layer — forcing Freehand to reveal whether its own pricing survives year two. Finally, the founders’ pattern — leaving Series B-funded Pando to rebuild its best module next door — is a diligence talking point competitors will happily circulate.

Same playbook, new buyer

Freehand hunts the Fortune 500 — Meta, Pfizer, Unilever — where deals are bespoke and sales cycles long. The same Category Context Graph mechanics work down-market: mid-market manufacturers and distributors ($100M-$1B revenue) leak freight and supplier spend at proportionally higher rates because they never had the BPO armies Freehand displaces, and they need self-serve, priced-on-a-page products rather than negotiated enterprise contracts — a motion Freehand’s white-glove model can’t serve without rebuilding its go-to-market. Second shift: sell to the intermediaries Freehand explicitly bypasses. Battery’s Thakker framed the positioning as targeting shippers over freight intermediaries, which leaves 3PLs, brokers, and forwarders — who process invoice volume on thin margins and would pay for agentic audit as an internal cost weapon — as an open channel someone like Loop is only partially covering. Third: adjacent spend categories with identical contract-versus-invoice mechanics — healthcare claims-versus-contract auditing, construction subcontractor payment applications, utilities and telecom spend — where the audit playbook transfers but Freehand’s supply-chain-specific graph and brand do not. The incumbent won’t chase these because its entire valuation story is Fortune 500 supply-chain depth; going broad or down-market would dilute the exact positioning its Series B was priced on.

Sources and further reading

• Freehand Raises $75M Series B To Automate Fortune 500 Supply Chain Spend (Crunchbase News, Jul 29, 2026)
• Freehand Raises $75M to Scale AI Teams Managing Supply Chain Spend for Fortune 500 Companies (GlobeNewswire / Freehand, Jul 29, 2026)
• Freehand Debuts as New Company with Dedicated AI Teams for Supply Chain Spend Management (Yahoo Finance / press release, Feb 2026)
• Freehand Raises $75 Million As Its AI Agents Target The Supply Chain (Forbes, Jul 29, 2026)
• AI-powered supply chain startup Pando lands $30M investment (TechCrunch, May 2023)
• Chennai startup Pando raises Rs 64 Cr Series A led by Chiratae Ventures (YourStory, Jan 2020)
• AI Freight Audit and Payment Agents (Freehand company site, accessed Jul 2026)
• Levelpath Raises $55+ Million to Reimagine Enterprise Procurement, led by Battery Ventures (BusinessWire, Jun 2025)
• Loop and J.P. Morgan Payments to Improve Access to Working Capital Across the Supply Chain (BusinessWire, Jun 2024)
• Freight Audit And Payment Market Size (Mordor Intelligence, 2025)
• Pando Reviews (Glassdoor, accessed Jul 2026)