Diligent x Forest: the Fincrime/AML agent, under governance

Diligent automates compliance, Forest keeps the proof

When regulators fine a bank or a financial institution, the industry’s answer is still the same: more headcount.

But when a fast-growing fintech or bank scales into millions of transactions, millions of active customers and millions of cross-border payments, headcount stops keeping up. At that volume, no compliance team can give every high-risk case the attention it deserves. The math doesn’t work on headcount alone.

Diligent takes that volume off their plate. Diligent’s agentic solution investigates, runs onboarding risk checks and clears false positives the way an analyst would, working on top of the screening providers the team already uses. So the compliance team can spend its judgment where it counts: the genuinely high-risk cases.

Forest holds the record. Every check, decision and piece of supporting evidence lands in a single case file, ready for an auditor or a regulator without anyone rebuilding the trail after the fact.

Automated where it should be, documented where it counts.

Who Diligent is

Diligent AI builds autonomous agents for financial crime compliance. The goal is to enhance AML ops teams’ capabilities by automating reasoning-heavy work: due diligence and name screening that financial institutions run every day.

Raised $3M, Diligent is YC-backed, and was founded by Edoardo Maschio (CEO) and Ahmed Gaber (CTO) alongside a superstar team (ex-Meta, Amazon, Billie, Citi, BCG), all of whom are ex-founders or part of founding teams.

Diligent’s AI agents sit on top of the tools you already use and resolve the work end-to-end. Two are most relevant here:

• AML Screening agent: natively integrates with your screening providers (e.g. LexisNexis, World-Check, ComplyAdvantage, and Dow Jones), investigates and clears false-positive sanctions, PEP, and adverse-media alerts.

• Merchant risk agent: reviews a business’s risk from its own web presence, socials, reviews, and registry filings, catching risky merchants at onboarding.

Compliance teams no longer rip out and replace their tech stack. Diligent adds an intelligence layer on top of the tools you already run, so integration is seamless and the implementation cycle drops to almost zero. Forest is the access and audit layer it runs inside, so every decision that closes an alert is logged like any other action.

Reduce manual alert handling, at 99.9% precision

Flywire is a global payments platform, moving cross-border payments across education, healthcare, travel and B2B in dozens of countries. At that scale, millions of cross-border payments generate a sea of name screening alerts, and the backlog waiting on analysts grows faster than any team can clear it, putting the whole AML operation under strain.

Flywire entrusted its name screening procedure to Diligent. The AI agent natively sits on top of their existing screening provider, investigates each alert, adds commentaries, identifies the false positives and closes them with no human in the loop. And Flywire trusted Diligent with the strictest, highest-stakes flow of all: sanctions screening.

Cross-border volume means a constant stream of hits, and almost all of them are namesakes: a payer who happens to share a name with someone on a sanctions list. This is where cultural nuance decides the outcome. Diligent reads names with that knowledge, catching the mismatches a rules engine treats as a match, and works from the little information a payment message carries, parsing the message itself to enrich the alert and confirm the false positive. To date, Diligent cuts Flywire’s manual alert handling by 50% while holding 99.9% precision on the alerts closed as false positives.

One question remains: when an agent reads a screening hit, decides a match is a false positive, or clears a sanctions alert on its own, who guarantees it did only what it was allowed to do? And who can prove it to the regulator six months later?

An agent governed like an analyst

That’s exactly where Forest comes in. An agent acting in a regulated environment raises the same questions as a human operator, only sharper: what rights, over what data, with what trace.

Forest runs Diligent’s agents inside its runtime, under the same permission model and the same audit log as humans. The agent has a role, like an analyst. It sees only the fields its role allows, triggers only the actions it’s permitted, and every decision it makes, closing an alert as a false positive, escalating a hit, writing a disposition, leaves a record-level trace, with the reasoning that produced it. An analyst and an agent held to the same rule, the same approval gate, the same log.

In practice, Diligent’s name screening agent plugs in as a call inside a workflow, sanctions, PEP, adverse media or ongoing monitoring, orchestrated on Forest. It returns its assessment on each hit, and Forest decides what happens next according to rules the compliance team configures: the alert closes as a false positive, or it’s sent for human review, with the agent’s analysis already attached. And the format of that audit trail, the one a regulator like the ACPR will want to see, is co-built so the agent’s decision is defensible, not just fast.

That’s the direction Forest is heading, and one we already deploy with clients through partners like Diligent: agents that take on the routine, an infrastructure that keeps control and proof. Diligent and Forest are working together on this model, with pilots underway.

An agent is only worth what you allow it to do

Diligent solves the investigative capacity problem: agents that take on the routine screening work, so teams can put the love and care into truly high-risk cases. But what’s missing is governance of the workflow, and of an agent that accesses more context than it needs.

That’s Forest’s job: giving these agents a role, limits, and an auditable memory, exactly like a human colleague. The agent does the work. The infrastructure guarantees it did so within the rules, and keeps what it takes to prove it.

If you’re evaluating compliance agents and your risk team’s first question is “sure, but how do we audit it,” that’s the exact type of conversation we have every day. Book time to share with us the screening process you’re trying to automate.