SAR Agents, wider payments screening, and everything else we shipped in July
This was a good month for anyone who files SARs or screens payments for a living. SAR Agents now draft a FinCEN-ready narrative straight into a filing. Payments Screening picked up PEP and Adverse Media watchlists on top of sanctions. goAML moved onto the same filing form generator as SAR, auto-assignment learned to route by entity and instrument, and matchlists, SAR filing consolidation, and alert exports all got upgrades of their own. Here's what's new.
SAR Agents draft your narrative for you
A SAR narrative has to pull together everything tied to a filing: alert dispositions, case notes, entity profiles, transaction history, and comments from every analyst who touched the investigation along the way. Someone still has to turn all of that into the five W's and how (who, what, when, where, why, and how) in a form FinCEN and examiners will actually accept.
SAR Agents do that synthesis now. It's a new AI Agent built to run on the SAR filing itself, pulling in every associated case and alert, the narratives already written for each, the transactions and entities involved, and the notes attached along the way. From that full picture, it drafts a FinCEN-compliant narrative straight into the filing's narrative field, shaped by whatever Narrative Instructions your team has set up, with FinCEN requirements always taking precedence over those. Every draft gets versioned in the SAR's audit trail, and an analyst reviews it before anything goes out the door.
What's nice about it is what it does to the narrative itself. One built from the complete filing, not from whatever an analyst can piece together from memory under a deadline, gives examiners a fuller picture right from the first draft. It also frees up the analyst's time for the part that actually needs a human: does this narrative hold up, is anything missing, what deserves a second look. Less time spent rebuilding context that already lives in the system.
SAR Agents pick up where Case Agent left off. Case Agent synthesizes everything tied to a case into a narrative the moment an analyst opens it. A filing spans multiple cases, so this is that same idea, just one level up.
Payments Screening now covers PEP and Adverse Media watchlists
Payments Screening, our real-time transaction screening, now checks activity against PEP (politically exposed persons) and Adverse Media watchlists on top of the sanctions lists it already screens against, like OFAC's SDN list. All three run in real-time, on every screened transaction.
That's a fuller risk picture at the exact moment a transaction happens, not after the fact in a batch review. If you're already a customer, talk to your CSM about adding PEP and Adverse Media watchlists to your configuration. New customers choose the watchlists they want during onboarding.
goAML filing gets the same form generator as SAR
goAML filings now run on the same shared form generator that powers SAR and our other regulatory filings. You still get full UNODC schema compliance under the hood, but the filing experience now matches the workflow analysts already use for every other filing type in Unit21.
If your team files goAML under UNODC standards across different Financial Intelligence Units, that consistency is really the point. One filing workflow to learn, not a different one for every report type.
Smarter routing for alert auto-assignment
Auto-assignment schedules can now group alerts by the flagged entity or flagged instrument, alongside the round-robin, creation-order option that's always been there. Turn grouping on, and an alert routes to whoever already reviewed a previous alert tied to that same entity or instrument. The analyst who already has the context gets the next one too.
Matchlists now show up across Agentic Task Builder, alerts, cases, and filings
Matchlists, the lists teams maintain of known entities, blocked businesses, watchlisted counterparties, are now a supported data source in Agentic Task Builder, across both its Online Search and Data Analysis builders. Hits surface directly on alerts, cases, and filings, and investigators can query matchlist data inside a custom Data Analysis task the same way they'd query any other data source, right alongside the objects they already work in every day.
Two more worth knowing about
Add Cases to Existing SAR Filing. Related cases don't always turn up before a SAR is already underway. Now they can be folded into that same filing instead of starting a new one from scratch.
Watchlist results on alert exports. Alert exports now include a full watchlist screening sheet, with match scores, PEP info, categories, and more, so exported data lines up with what analysts already see in the platform.
That's July, wrapped. A handful of releases that add up to real-time back for the people using them every day. Thanks for reading, and see you next month.
Want to see the AI Agents in this post in action? Register for See it live: one AI layer across detection, investigation, and filing, a live walkthrough of rule recommendations, alert and case review, SAR filing, and custom tasks you build yourself.
Kunal Datta is the Chief Product Officer at Unit21. Prior to Unit21, he led the Product team for Checkout at Fast, and prior to that, led the Product teams responsible for automating aerial wildfire safety inspections at Pacific Gas & Electric.
He has a background leading Product teams using AI to automate processes at regulated entities, as well as financial products, machine learning products, web applications, mobile applications, hardware products, and data products. Kunal is a Fulbright Scholar and studied Civil and Environmental Engineering and Music Science Technology at Stanford University.
