Footprint Raises $25M for Auditable Financial-Crime AI

Financial-crime investigations create a second product after the alert is cleared: precedent. A bank needs to know why an analyst approved, escalated or dismissed a case, whether the same logic holds across KYC, sanctions and transaction monitoring, and whether an examiner can reconstruct the decision months later.

Footprint is building around that problem. The New York financial-technology company announced a $25M Series B led by QED Investors to expand Percy, its agentic operating system for financial-crime compliance, and Trust Fabric, the governed institutional memory beneath it. MUFG, Commerce Ventures, LightBank and Alumni Ventures participated alongside returning investors Index Ventures, Lerer Hippeau, BoxGroup, Operator Partners and Animal Capital.

From alert automation to organizational memory

Compliance teams already use software to collect data, screen names and create alerts. The harder work begins after a system finds something: an analyst must gather context, apply policy, distinguish a real risk from a false positive and leave a record another person can understand.

Footprint’s thesis is that AI should not merely return an answer. It should carry forward the evidence, policy and institutional decisions that shaped that answer. Percy’s agents can investigate cases and propose actions across KYC and KYB, sanctions and watchlist screening, transaction monitoring, enhanced due diligence and adverse media. Humans approve the action, while citations, timestamps and task history remain available for review.

Trust Fabric is intended to make those decisions reusable. If one analyst resolves a difficult match or establishes how a policy applies to a particular fact pattern, the system can preserve that context for the next related case. That shifts the product from task automation toward organizational memory.

The distinction matters because financial-crime teams face both volume and accountability. Nasdaq Verafin’s 2025 global report estimated that illicit financial activity reached $4.4 trillion in 2023 and found growing concern about AI-enabled attacks. Institutions want more automation, but regulators and internal risk leaders still need to understand how decisions were reached.

The performance claim includes an audit trail

Footprint says its software has reduced some watchlist reviews from 30 minutes to less than one minute. It also says enhanced-due-diligence reviews that took three hours can be completed in under 15 minutes while surfacing 35% more evidence. Its product page separately advertises 20-times-faster reviews, four times more data sources, full auditability and automation of 90% of level-one reviews.

These are company-reported figures, not independently audited benchmarks. They nevertheless show how Footprint is positioning the product: speed is valuable only when the institution retains a defensible decision record. The company says customers include FDIC- and OCC-regulated banks as well as Bilt, Nuvei and MoonPay.

The human-approval design is also central to the pitch. Agents can gather information, compare it with policy and recommend an action, but a person remains responsible for the consequential decision. That structure may help banks introduce AI into regulated workflows without treating the model as an unaccountable final authority.

What the Series B is meant to fund

Footprint plans to use the capital to double its engineering and sales organizations, continue research and development for Percy and Trust Fabric, and open a San Francisco office alongside its New York headquarters. Co-founder and CEO Eli Wachs described the round as bringing Footprint’s total funding to $45M. Co-founder Alex Grinman serves as CTO.

The funding history deserves a small qualification. Footprint publicly announced a $6M seed round led by Index Ventures in 2022 and a $13M Series A led by QED in 2024. A 2024 SEC Form D records $20.028M sold from a $20.16M offering, which does not map cleanly to the announced $13M Series A. The company has not publicly explained that difference, so the $45M total should be treated as management’s stated figure rather than a reconstructed total from disclosed rounds.

Terms beyond the round size were not disclosed. Footprint did not announce a valuation, ownership changes or a new board appointment, and it did not disclose revenue, profitability, pricing, retention, customer concentration or exact hiring targets.

The governance question moves to the foreground

Institutional memory can compound good judgment, but it can also preserve an old mistake. A system that reuses prior decisions needs controls for contradictory precedents, changing policies, data access, reviewer authority and the retirement of reasoning that is no longer valid. As agents take on more investigative work, banks will also need to decide which actions always require a person and what evidence is sufficient for approval.

That tension is the most important part of Footprint’s bet. The company is not only selling faster alert resolution. It is arguing that auditability, evidence and governed memory are the features that make AI usable in compliance at all.

The Series B gives Footprint more capital to prove that argument with regulated institutions. The next measure of progress will be whether customers can expand automation without losing the clarity of ownership that regulators, auditors and risk teams expect.