A New Era. A New Unit21 ๐
The math of risk and compliance is broken.
Transaction volumes have exploded. Alert queues are overwhelmed. And traditional systems mostly create noise. โMore of the sameโ is no longer a strategy.
A new era of financial crime requires a new architecture to defend against it.
Today, weโre introducing a new Unit21: AI Risk Infrastructure
The Shift
Passive monitoring tools are no longer enough. Most systems simply record the workโthey donโt actually execute investigations end to end. They donโt automatically tune your system for better fincrime detection. Until now.
The AI Evolution
AI isnโt just another feature. Itโs a new operating model.
Weโve built AI Risk Infrastructure that acts as a force multiplier for financial crime teams.
AI Agents that:
๐ฃ๐ฒ๐ฟ๐ณ๐ผ๐ฟ๐บ ๐๐ต๐ฒ ๐๐ผ๐ฟ๐ธ โ eliminating manual data hunting
๐ฆ๐ต๐ผ๐ ๐๐ต๐ฒ ๐๐ผ๐ฟ๐ธ โ providing instant, transparent context
๐๐บ๐ฝ๐ผ๐๐ฒ๐ฟ ๐๐ฒ๐ฎ๐บ๐ โ delivering finished investigations, not just alerts
The future of fraud prevention and AML isnโt about how many alerts you generate. Itโs about how much risk you resolveโwith precision and automation.
A new era is here.
Read more from our COO, Tyler Allen: https://t.co/Wb6pNCpmeg
Transaction monitoring. Market abuse rules. STR filings. Regulator expectations that keep evolving.
MiCA authorized is not the same as MiCA ready.
On June 10, we're bringing together two practitioners who are living this right now for a candid panel discussion on what it actually takes to operationalize MiCA compliance at scale.
Joining the conversation:
โณ Eoin Kearns, Head of Compliance & AML, @moonpay โ led MiCA compliance from inception, launched across the EEA, and has spent the last year working directly with the AFM to uphold their commitments
โณ Calley Jansen, Head of Financial Crime Intelligence, @BVNKFinance โ on the operational front lines of MiCA: transaction monitoring, market abuse rules, alert efficacy, and governance
Moderated by Emily Garza, Chief Customer Officer, Unit21.
No legal explainers. Just practitioners who've built it, talking honestly about what works.
๐ June 10 | Register โ https://t.co/cOKi0pO9eL
Unit21 quietly files roughly 5% of every SAR in America. Most banks have never heard of them.
Our cover story on @unit21inc
https://t.co/jUgGqprwC9 via Yespress
The hardest part about AI agents making payments is not the transaction itself. Itโs proving intent, identity, and liability at the same time when there is no human directly pressing the button anymore.
What Tyler said during our conversation stuck with me because the industry is entering a world where fraud detection is no longer just about checking passwords, OTPs, or suspicious IP addresses. AI agents will operate from devices, behave like humans, and in many cases even imitate normal customer patterns better than fraudsters ever could.
That changes the entire risk model.
An AI agent making a payment for me might actually look legitimate because it knows my normal behavior, transaction sizes, merchants, and habits. So the decision engine starts relying less on one signal and more on layered intelligence across device behavior, transaction history, behavioral patterns, and agent detection.
The interesting part is that โthis is an AI agentโ does not automatically mean โthis is fraud.โ
It simply becomes another signal in the flow.
@unit21inc
Unit21 was just named a Category Leader by @ChartisRsch and ranked #1 in AI across 40+ fraud vendors.
We earned best-in-class scores in workflow, modeling, case management, and configurability.
If you're evaluating fraud platforms in 2026, this report is worth reading. It maps the market and explains what separates the leaders from the rest.
Download full report๐https://t.co/KfkB7pIYFx
Banks are still fighting AI fraud with yesterday's playbook.
Our CEO Tyler Allen sat down with @samboboev from Fintech Wrap Up Podcast to talk about why that's a problem - and what it actually takes to fix it.
In this podcast, he covers:
โณ Why rules-based systems and ML models aren't enough anymore
โณ How AI agents are changing fraud and AML operations
โณ The future of Know Your Agent and regulatory readiness
โณ Why compliance teams may go AI-native faster than expected
Worth a listen ๐ง โ https://t.co/A7cMg9P7Uw
Fraud teams used to spend months building detection logic, tuning rules, and manually reviewing edge cases, but Tyler explained how fast that workflow is changing when AI can now generate and deploy fraud rules from a simple text prompt instead of requiring analysts to understand data models, SQL queries, or complex dashboards.
What stood out to me is that the future of fraud infrastructure may not even look like software people actively use every day. The system quietly runs in the background, monitors behavior, deploys detection logic, flags anomalies, learns from feedback, and only pulls humans in when something truly unusual happens.
That changes the role of fraud teams completely.
Instead of spending hours configuring systems, teams start focusing on oversight, exceptions, and strategy while AI handles repetitive operational work at scale. A company processing millions of transactions per day cannot realistically rely on manual reviews and static rules anymore when fraud patterns shift weekly and attackers already use AI themselves.
The interesting part is that fraud software is slowly becoming less of a dashboard and more of an invisible operating layer for compliance, risk, and decision-making.
@unit21inc
Tyler made an interesting point on the podcast when we discussed AI in fraud protection.
Last year, many financial institutions were still treating AI agents as experiments inside innovation labs. Now more than 50% of financial services companies are already deploying them or actively trying to move into production environments as quickly as possible.
What changed is simple. Fraudsters are already using AI at scale. They can launch thousands of synthetic identity attempts, phishing campaigns, or account takeover attacks in minutes, while traditional fraud teams still rely on workflows built for a slower internet.
The pressure is no longer about whether companies should adopt AI for fraud prevention. The pressure is whether they can adapt fast enough before attackers widen the gap even further.
What also stood out from Tylerโs point was how quickly the market sentiment changed in just 12 months. Moving from roughly 10โ20% adoption to the majority of organizations exploring deployment shows how fast financial infrastructure decisions are now shifting once operational risk becomes impossible to ignore.
@unit21inc
๐ฌ๐ผ๐๐ฟ ๐ฐ๐ผ๐บ๐ฝ๐น๐ถ๐ฎ๐ป๐ฐ๐ฒ ๐ฝ๐ฟ๐ผ๐ด๐ฟ๐ฎ๐บ ๐ฏ๐๐ถ๐น๐ ๐ฎ ๐๐ฅ๐ฅ ๐บ๐ผ๐ฑ๐ฒ๐น.
Weighted attributes. Configured thresholds. A score for every customer. And when an investigator opens an alert, they see: High. 82/100. That's it.
They don't see that SAR_VOLUME is carrying 20% of the model weight and scored 100/100 because this customer filed two SARs in the last 12 months.
They don't see that AGE_OF_RECORD scored 0/100 because a 66-month-old account is considered low-risk by your model's configuration.
They see a number. They don't see the reasoning behind it.
So they can't document it. They can't challenge it. And when an examiner asks why a case was escalated, "the CRR was high" isn't a sufficient answer.
๐๐ฉ๐ช๐ด ๐ช๐ด ๐ต๐ฉ๐ฆ ๐ฑ๐ณ๐ฐ๐ฃ๐ญ๐ฆ๐ฎ ๐ฐ๐ถ๐ณ ๐๐ถ๐ด๐ต๐ฐ๐ฎ๐ฆ๐ณ ๐๐ช๐ด๐ฌ ๐๐ข๐ต๐ช๐ฏ๐จ ๐๐ฏ๐ข๐ญ๐บ๐ด๐ช๐ด ๐ต๐ข๐ด๐ฌ ๐ด๐ฐ๐ญ๐ท๐ฆ๐ด.
The moment an investigator opens a flagged alert or case, the task automatically surfaces the full attribute-level breakdown: each factor's weight in the model, its individual score, the underlying metric, the raw entity value, and the mapped score. The complete logic chain, already there, before the investigation begins.
๐ ๐๐ฐ๐ผ๐ฟ๐ฒ ๐๐ฒ๐น๐น๐ ๐๐ผ๐ ๐๐ต๐ฎ๐. ๐ง๐ต๐ฒ ๐ฏ๐ฟ๐ฒ๐ฎ๐ธ๐ฑ๐ผ๐๐ป ๐๐ฒ๐น๐น๐ ๐๐ผ๐ ๐๐ต๐.
Edition 03 of AI Task Spotlight is live. Read how it works:ย https://t.co/FmHz8LflHH
#AML #Compliance #CustomerRisk #TransactionMonitoring
New email. Same questions. Wait. Chase. Screenshot the thread. Upload it manually. Log it in a spreadsheet.
Every compliance analyst knows the drill.
That's not an audit trail. That's just hoping nothing goes wrong before the exam.
We built Request for Information (RFI) to fix it.
From any alert or case in Unit21, your team sends a secure, white-labeled form, sets a deadline, and receives everything back in the platform automatically. Full audit trail. No chasing. No manual uploads.
RFI is live. Read more: https://t.co/6mhWEnSpRz
Tyler made an important point during our conversation that fraud teams are now operating in a completely different environment because AI gives attackers an unfair advantage in scale, speed, and experimentation.
A fraudster only needs one successful hit out of 100 attempts to make money. A bank or fintech cannot afford that same error rate because one hallucinated decline can block a legitimate customer from accessing financial services, trigger compliance issues, or destroy trust instantly.
That is what makes this shift so difficult for financial institutions.
The same AI acceleration helping fraudsters is also giving risk and compliance teams entirely new capabilities. Real-time behavioral analysis, adaptive onboarding checks, AI-driven monitoring, and autonomous investigation systems are moving from experiments into production infrastructure much faster than most people realize.
The question is no longer whether AI can support fraud and compliance operations.
The real challenge is how quickly regulators and financial institutions become comfortable allowing AI systems to participate directly in critical risk decisions while still maintaining accountability, explainability, and accuracy at scale.
@unit21inc
โฐThis Thursday, nearly 800 compliance leaders are tuning into this conversationโฆ
FinCEN's NPRM is reshaping what AML program effectiveness actually means - and the comment period closes June 9.
Before it does, join us and American Banker for a live, unfiltered conversation on what compliance leaders need to know.
Joining the conversation:
โณ Sarah Beth Felix, Palmera Consulting - recovering auditor, founder, and AML consultant who's seen this from every angle
โณ Eric Ellis, @FifthThird Bank - former OCC BSA Policy Director who's sat on both sides of the exam table
Spots are still open.
๐ May 28 | 1:00 PM ET
Register โ https://t.co/d5kTBjS8Gr
One thing that keeps coming back in conversations around AI agents and payments is that the industry is still thinking mostly about capability, while the harder problem is governance.
An AI agent being able to execute a payment is not the difficult part anymore. The infrastructure already exists. APIs exist. Stablecoins exist. Real-time payment rails exist.
The real challenge starts when an agent makes a mistake, exceeds permissions, gets compromised, or negotiates on behalf of a customer and moves money autonomously. At that point, the question becomes very simple: who is responsible for that action?
That is why Tylerโs point about an โAI agent registryโ is so interesting.
Not because registries are exciting technology, but because financial systems have always depended on trusted identity and accountability layers sitting underneath transactions. Merchants have IDs. Banks have licenses. Payment providers have scheme memberships.
AI agents currently have almost none of that structure.
I would not be surprised if companies like Visa, Mastercard, or Stripe eventually become part of this layer because they already sit in the middle of trust, compliance, permissions, authentication, and dispute management across global commerce.
@unit21inc
"Unit21 gives us the flexibility and control to move fast. We can tailor rules, test safely, and scale compliance as quickly as Rippling grows."
Their old compliance stack wasn't built for scale but with Unit21, they got real-time transaction blocking before funds settle, dynamic customer risk scoring, and a single workflow from alert to SAR filing.
Read the full case study โ https://t.co/ptYqT2AiAV
In this episode of the WRAP UP Podcast, I speak with Tyler Allen, CEO of @unit21inc, about why AI agents may become one of the most important tools in the fight against financial crime.
Tyler explains why traditional rules-based systems and machine learning models are no longer enough as fraud becomes faster, more automated, and more sophisticated. He also breaks down how AI agents can help compliance and fraud teams review alerts, investigate suspicious activity, reduce false positives, and respond to financial crime in real time.
The conversation also explores the future of โKnow Your Agent,โ AI-powered fraud decisioning, agent registries, regulatory readiness, and whether financial institutions are ready to trust AI with high-stakes compliance work.
In this episode, we discuss:
- Why financial crime is becoming harder to fight
- How AI is changing fraud and AML operations
- Why traditional machine learning is not enough
- The role of human-in-the-loop reviews
- How AI agents can reduce investigation time
- The future of Know Your Agent
- Whether regulators are ready for AI decision-making
- Why compliance teams may become AI-native faster than expected
Watch the full episode to understand how AI agents could reshape fraud prevention, AML, and financial crime compliance.
The comment period for FinCEN's NPRM closes June 9.
Before it does, we think compliance leaders should understand exactly what's changing, what's cosmetic, and what will actually move the needle in an exam room.
We're joining @AmerBanker for a live conversation on May 28 to break it all down.
Joining the conversation:
โณ Eric Ellis, @FifthThird Bank
โณ Sarah Beth Felix, Palmera Consulting
โณ Neda Shapourian, Unit21
On the agenda: what "effectiveness" actually means to an examiner, how AI earns regulatory credit vs. just providing cover, and how small teams can make defensible risk-based decisions.
Real talk only.
๐ May 28 | PM ET
Register โ https://t.co/d5kTBjS8Gr
Most compliance teams think they have watchlist screening covered.
They have ongoing monitoring. Onboarding checks. Maybe a manual process for one-offs.
What they actually have is three vendors, three interfaces, and three audit trails, and when a regulator asks for a consolidated view, they're piecing it together by hand.
That's not a program. That's a patchwork.
A complete screening program covers five moments in the customer lifecycle: onboarding, ongoing monitoring, ad hoc checks, payment screening, and AI-powered investigation. All five. One platform. One audit trail.
We broke down exactly what that looks like, plus a full demo walkthrough so you can see it end-to-end.
๐ Read the blog โ https://t.co/YYpBqQC4Nx
The Nacha 2026 rule changes arenโt just about detection; theyโre also largely about documentation.
Investigators can't audit what isn't logged. Examiners don't just look at whether your rules fired. They look at your rule inventory, modification history, performance data, and whether your program shows evidence of continuous improvement.
We just published Part 3 of our Nacha 2026 series breaking down exactly what explainable and auditable documentation looks like and how to build it.
If your team is preparing for the annual assessment, this one's worth a read.
๐https://t.co/URj4i8dtRy
A lot of great conversations these past few weeks.
From @ABABankers Risk & Compliance in Charlotte, to @acams_aml Kansas City, to @qubevents Payments & Regtech in Barcelona, and Corelation Client Conference in San Diego, one thing kept coming up in every room: financial crime teams are being pushed to move faster without losing control.
AI, regulatory pressure, operational scale, cross-border complexity. The challenges may look different depending on the audience, but the direction is pretty consistent.
Thank you to those who spent time with us across sessions, booth meetings, and conversations over cocktails. Special shout out to our friends at @Equifax for co-hosting our mixer in Charlotte!
A few highlights from the 4 events spanning 7 days and over 6000 miles ๐ธ
You got your MiCA license. Now the real work starts.
We've found that most MiCA conversations stop at licensing, so weโre bringing together operators from @moonpay, @BVNKFinance, and @flowdesk_co to talk about ๐๐ต๐ฎ๐ ๐ต๐ฎ๐ฝ๐ฝ๐ฒ๐ป๐ ๐ฎ๐ณ๐๐ฒ๐ฟ ๐ฎ๐ฝ๐ฝ๐ฟ๐ผ๐๐ฎ๐น.
This will be a practical conversation about what teams are seeing on the ground, whatโs changing under MiCA, and where companies are still getting stuck.
๐ Register here: https://t.co/vSAlCSRT7C