What shipped in June: the case investigates itself before you open it
Consider the first ten minutes an analyst spends on a case. Before any real thinking happens, there is assembly. Open the alert. Open the entity profile. Check adverse media. Trace the network connections. Copy a little from each tab into a working narrative. Only then does the actual investigation begin.
June was largely about removing that first ten minutes, and a good deal of the friction that surrounds it. We shipped a new way for AI to investigate a whole case, a way for customers to plug their own AI agents directly into Unit21, a new filing type, a way to consolidate SAR filings, and a handful of upgrades across case management, real-time rules, and risk scoring. Here is what is new.
Launch spotlight: Case Agent is live
Our AI agents now investigate the whole case, not just the alert.
Every attached alert, entity, adverse media hit, and network connection gets synthesized into one narrative before an analyst opens the case. The narrative is ready and waiting the moment a case is opened, with no manual pulling together of context across tabs and data sources.
Why it matters is best seen by comparing it to the way the work happens today. Case investigation means stitching together alerts, entity profiles, adverse media, and network data by hand before the real analysis can even start. Case Agent does that stitching automatically. An analyst opens a case to a synthesized narrative instead of a pile of raw data, and can spend their time on judgment instead of assembly. If you want to see how this fits into the investigation flow, our case management page walks through it.
AI Agent API Suite is live
For a while now, customers building their own AI tooling have been asking for a clean way to connect directly to Unit21. This is that path.
Customers can register their own AI agents directly in Unit21, assign them tasks from a catalog of Unit21 capabilities, kick off runs, and track results in real time. The suite is standalone and production-ready today, so customers can plug their own AI tools into Unit21 without going through the UI. And there is full visibility into what agents are doing and how runs are progressing at every step.
The API Suite gives those teams a stable, documented path to build on today. It is also the starting point for deeper agent-to-agent integration, rather than the finish line.
Two upgrades for the filing workflow
Two of June's launches sit squarely in the compliance filing path.
First, DOEP (Designation of Exempt Person) filing. FIs can now file a DOEP with FinCEN directly in Unit21 to exempt qualifying customers, such as other depository institutions, government entities, listed public companies, and eligible high-volume cash businesses, from CTR filing requirements on cash transactions over $10,000. DOEP is a one-time filing that removes ongoing CTR burden for qualifying customers. Support for it closes a gap for FIs with high-volume cash activity, and it was one of the fastest filings the team has shipped to date.
Second, adding cases to an existing SAR filing. Users can now add additional cases to an existing SAR filing directly from Summary > Cases. Consider how investigations actually unfold. Related activity often surfaces after a SAR is already underway. This lets teams consolidate that activity into a single filing instead of starting a new one from scratch.
Quality-of-life across the platform
The rest of June was spent on the small repetitions that add up over a day.
Mixed event subtype variables in Real-Time Rules and Flow nodes. Real-time rules and flow nodes can now use variables that aggregate on event types different from the incoming event type being evaluated. Many real fraud patterns span event types, so this removes a modeling constraint that previously forced teams to work around mismatched event types when building velocity rules.
RTR Rule Silencing. Users can silence real-time rules for receiving entities directly from the rules page, entity details page, or via action buttons on the alerts page. Silencing to allowlist known-good receiving entities reduces repetitive false positives, so teams focus on genuinely suspicious activity. Silencing for sending entities is planned as a follow-up.
CRR Historic Risk Scores. A new timeline view and history table show how an entity's risk score has changed over time, with drill-down into what drove each shift, available on the entity sidebar and entity details Risk Scores tab. Risk scores change, but until now there was no easy way to see the trajectory or understand what moved it. Compliance teams now have that context in one place.
FFIP Enrichment. A new matching engine connects X9 check images and NACHA data to the correct core transactions in Unit21, using configurable waterfall rules and transformations instead of custom code. Check images and ACH addenda attach directly to the matched transaction, and every stream includes a Test Match step plus full audit lineage before anything goes live. Check and ACH data have historically been hard to tie back to the transactions they belong to. FFIP Enrichment closes that gap, so investigators get complete records without manual cross-referencing.
What's next
A theme runs through most of this. The platform is doing more of the assembly, the consolidation, and the matching that used to fall to a person. Notice that none of it removes the analyst from the decision. It removes the work that stands between the analyst and the decision. If you want the longer arc of where this is heading, our 2025 product recap traces the same thread across the year.
More is coming soon, including Unit21 AI summer camp. And if you happen to be in Chicago, join us and TRM Labs for the Crypto Happy Hour on Thursday, July 30 at the LondonHouse Rooftop. Space is limited, so request an RSVP.
See it on your own cases
The fastest way to understand Case Agent and the AI Agent API Suite is to watch them run against the kind of activity your team sees every day. Book a demo and we will walk you through it. Existing customers, reach out to your CSM to turn any of the above on.
Gal Perelman is the Product Marketing Lead at Unit21, where she spearheads go-to-market strategies for AI-driven risk and compliance solutions. With over a decade of experience in the fintech and fraud sectors, she has led high-impact launches for products like Watchlist Screening and AI Rule Recommendations.
Previously, Gal held marketing leadership roles at Design Pickle, Sightfull, and Lusha. She holds a Master’s degree from American University and a Bachelor’s from UCLA, and is dedicated to helping banks and fintechs navigate complex regulatory landscapes through innovative technology.
