Parcha

The Future of Compliance is Intelligence ?? In the latest edition of The Audit Log, our quarterly deep dive into what we're building at Parcha, we unpack how Parcha's AI continues to transform how compliance teams work and our vision for a shared layer of intelligence layer across all compliance workflows. Here's what's in the update: ?? Individual Due Diligence 2.0: What happens when you apply the same AI that automates business reviews to researching individuals? Our early results show what used to take an hour can now be done in minutes. ?? Breaking the Document Fraud Arms Race: By studying how expert analysts catch sophisticated fraud, we've built a multi-modal AI that doesn't just read documents - it sees them. The results? 3x more fraud caught with zero increase in false positives. ?? Smarter AML Screening: We're moving beyond simple name matching to understand the full context of each alert helping teams cut through the noise of false positives while maintaining the highest compliance standards. ?? The Future of Compliance: We share our vision for a unified System of Intelligence for Compliance - an AI layer that works across every workflow and integration to make compliance teams more effective. Want the full story? Check out the link in comments!

Why is traditional AML screening fundamentally broken? ?? Last week, a fintech's head of compliance shared a jarring statistic: their team reviews tens of thousands of articles monthly for adverse media screening. But the real problem isn't just volume - it's about accurately identifying risk without drowning teams in false positives. Let's break down why current approaches fail: 1?? The Scale Problem A typical fintech screens 500+ new customers daily, plus quarterly re-screening of 25% of their customer base. 30% of accounts get flagged for adverse media, and each screening triggers 2-3 alerts with 10-20 adverse media hits each. That's thousands of articles daily that someone has to read. Manual review can't keep up. 2?? The Cultural Context Challenge Simple text matching can't understand that "José Ignacio Martínez" and "Nacho Martínez" might be the same person. These cultural and linguistic connections - like "Nacho" being a common nickname for "Ignacio" in Spanish-speaking cultures - completely escape traditional systems. Not to mention the fact that the adverse media articles are often in different languages. 3?? The Role Resolution Problem Consider this alert: "Nacho Martínez, CEO of Martinez Holdings, was defrauded in a large-scale investment scam." Traditional systems flag this as a hit because they found the name + "fraud" in proximity. But they miss crucial context - Martínez was the victim, not the perpetrator. Now multiply this challenge across thousands of alerts. 4?? LLMs Aren't All You Need While LLMs are powerful, they're just one piece of the puzzle. The real challenge is building intelligent pipelines with modular subsystems that can be independently evaluated and optimized. This architecture lets you use cheaper models strategically, leverage in-context learning where it makes sense, and maintain proper governance throughout. It's about intelligent systems, not just using frontier models. ?? Our Solution: Parcha's Intelligent AML Screening reduces global AML false positives by 90%. The impact? Compliance teams can finally escape manual review time waste and focus on what matters - identifying real risk. ?? Want to know how we do it? Learn in our latest technical deep dive in the link in comments...

Trump’s actions against Colombia over the weekend are a sign of more challenging times ahead for global AML compliance. A lot of folks have been assuming that financial crimes compliance will get easier under Trump’s administration but this weekend’s showdown with Colombia on immigration suggests otherwise. After the Colombian government reneged on their agreement to receive deported immigrants from the U.S., Trump threatened tariffs, travel bans and financial sanctions. Colombia quickly conceded but that doesn't mean other countries will do the same. While regulation may loosen, Trump’s use of sanctions as a threatening force to get countries to cooperate with US Foreign policy and immigration policy is going to make AML compliance more complex. Sanctioning countries over a weekend with little warning will create massive headaches for compliance folks as they have to quickly react to shutdown or offboard exposed customers. Furthermore, as sanctions list grow, false positives will rise. With the industry already suffering >90% false positive rates for sanctions screening, this is going to lead to even more wasted resources. We need a much more precise (low false positives) and dynamic way to manage and action on sanctions screening and that's exactly what we're building with Parcha’s System of Intelligence for Compliance. Our AI due diligence solution continuously builds context on an entity to reduce false positives while enabling compliance teams to act quickly to identify true positives when sanctions lists change. Do you feel ready for the shifting compliance landscape? If not, we're here to help. Reach out to learn more about Parcha! P.S. Check out the links in comments to our blog post and podcast episode from last year on how Trump’s second term might impact fintech and banking co

Driven by AI, we are entering a new era of enterprise software, ushering in systems of intelligence. In the mid 1980s, driven by the growth of the client/server architecture, we saw the dramatic rise of systems of record. These are the back office software applications that helped enterprises run their ERP, HR, CRM, and core IT workflows. These technologies were relatively specialized, and helped automate any of the most critical tasks in the enterprise. They were defined by structured data, back office automation, and leveraged by only by select users in an enterprise. With the rise of cloud and mobile in the mid 2000s, we saw a new era of systems of engagement, as coined by Geoffrey Moore. In a world of much more dynamic and ad-hoc work in the enterprise, systems of engagement were tools for collaboration, communication, video, work and project management, social and intranets, and more. These tools dealt with all the messy, unstructured data in an enterprise - the conversations, collaborative docs, and media that began to drive a shift in how the entire enterprise worked. Now, in the mid 2020s, we are firmly entering a new era of enterprise software, which gives rise to systems of intelligence. Systems of intelligence combine enterprise data, workflows, and AI, to deliver insights and automation to an organization. Importantly, because of the ability for AI to process unlimited unstructured data - like documents, video, or communications - we also get the same benefit from this messy data as we did our structured data. We can query, synthesize, calculate, and automate all the work around thus unstructured data just as easily as we could query a database before. Unlike systems of engagement that generally broke down the more information that goes into them, we see the reverse now with AI, where software can become more powerful and useful the more data it has access to. And with AI Agents being a native property of systems of intelligence, these systems aren’t only leveraged by every employee, they dramatically expand the output of the workforce. Systems of record are where people work by largely themselves. Systems of engagement let users work collaboratively with other people. Now systems of intelligence let us work seamlessly with people and AI. These systems will also talk to each other in completely new ways. Instead of deterministic APIs and clear handshakes, with Agentic AI, these systems will communicate with each other much like a humans do. A user will make a request in one system, and it will fan out the ask to a variety of other similar systems relevant for the desired information. And if it didn’t get what it wanted, it will simply request again in a different way, just as a person would. We’re going to see systems of intelligence in every domain of work - across every line of business and every vertical. Wild times ahead.

?? Why is fintech compliance becoming more fragmented just as risks are becoming more interconnected? A month after returning from Money2020, there's one thing I can't stop thinking about: As compliance tools proliferate, the ability for financial institutions to understand holistic customer risk is actually diminishing. Here's what I mean: Today's compliance teams are drowning in a sea of point solutions. One vendor is for KYC, another is for transaction monitoring, a third is for sanctions screening, and so on. Each tool generating its own alerts, maintaining its own customer data, and operating in isolation. The end result? A fragmented view of risk that's actually making compliance teams less effective, not more. This fragmentation creates three critical problems: 1?? Data Silos: Risk signals get trapped in individual systems, preventing teams from seeing the complete picture of customer risk. 2?? Workflow Inefficiency: Analysts waste precious time context-switching between tools and piecing together disparate information. 3?? Missed Connections: Important risk patterns go unnoticed because no single system can connect the dots across different types of compliance data. Other operational functions have solved this problem - think Salesforce for sales or Rippling/Workday for HR or ServiceNow for IT but compliance has remained stubbornly fragmented. Perhaps it's because compliance workflows are more complex or because the stakes around data accuracy are higher. There has to be a better way to solve this. Instead of adding more tools to the stack, what if we fundamentally rethought how compliance systems should work together? What if we could build a single source of truth that gives compliance teams a holistic view of customer risk while maintaining the specialized capabilities they need? It's a sentiment I'm hearing more and more from customers too: - "I don't understand why data from our AML alerts isn't feeding into our transaction reviews" - "We want one vendor that can solve all our needs" - "We need a customer lifecycle solution" I'm curious to hear from other compliance professionals: How are you dealing with vendor sprawl? What solutions have you found to connect the dots?

I'm super excited to welcome Xia to our founding engineering team! For this role we were looking for a hardcore builder who could work across the stack to solve any problem, including helping us scale our backend, jumping into applied AI when needed, and even working on frontend in a pinch. We were looking for the elusive 10X engineer, someone who can outbuild my co-founder Miguel Rios-Berrios. Here's what made us a "Hell yes!" on hiring Xia - He has a hardcore CS background and crushed our technical interviews - At Meta, he scaled Cassandra globally – fun fact I learned from Xia: Instagram is built on Python. Who says Python can't scale?! - At Brex he solved some gnarly problems in user roles and permissions - He spent the last year experimenting with AI and built a product to help realtors review lengthy contracts What made Xia stand out the most however, was that he approaches every challenge we throw at him with the words "Yes, let's do it." In his first week at Parcha, Xia shipped a demo end-to-end, including prompt engineering, for a new use case for one of our customers, which will soon convert to a six-figure contract! This was a great example of two of our company values: Act with urgency & Believe (like Ted Lasso) Welcome to the team Xia! -- If you want to spend less time on Zoom calls and more time shipping like Xia, DM me! We're hiring for applied AI and full-stack roles.

?? Did you know that before 1986, money laundering wasn't even a federal crime in the United States? ?? In 1970, a bank teller in Manhattan accepts a paper bag containing $50,000 in cash from a well-dressed customer. No questions asked. No forms filed. No eyebrows raised. This wasn't just common practice—it was perfectly legal. For criminals looking to launder money through the U.S. financial system, it was the golden age. The evolution of U.S. financial compliance is a fascinating story of how regulators adapted to emerging threats - from drug cartels in the 80s to internet banking and peer-to-peer payments in the 2000s to cryptocurrency today. ?? At Parcha, we've been researching the five landmark laws that transformed how financial institutions prevent crime and the evolution of AML legislation in the US over the last 50 years. ?? We've published an article and podcast breaking down how each law responded to the challenges of its time, and what this means for fintech builders today. For product managers, fintech and banking leaders trying to navigate an increasingly complex regulatory landscape, understanding this context is crucial. Check out our latest post and the episode of Compliance Accelerated (link in comments). Would love to hear your thoughts on which regulatory changes have most impacted how you build products. ?? https://lnkd.in/gVZxhMMN

One challenge keeps coming up in conversations with banks and fintech teams when it comes to business onboarding: the dreaded back-and-forth with collecting corporate documents for KYB. "We need one more document..." "Could you please resubmit this..." "This format isn't quite right..." ?? According to a recent study, 40% of consumers abandon banking applications due to these lengthy processes. ?? That's why I'm incredibly excited to share our latest product at Parcha: Real-Time Document Requests & Verification! Here's what makes this game-changing: ??♂? Transform weeks of back-and-forth into minutes ?? AI instantly verifies documents as they're uploaded ?? Immediate, clear feedback if something's wrong ?? One-click document requests for your team ?? Secure, personalized submission links In a recent proof of concept we completed with FIS, we achieved 98% accuracy on document verification without any additional customization. The best part? Watching customers smoothly complete their onboarding without the usual friction and frustration. ?? We're just getting started with reimagining the onboarding experience at Parcha. Link to the full blog post in the comments and don't forget to check out our other solutions at parcha.com

?? What Trump's Second Term Could Mean for Fintech, Banking and Crypto Compliance There's a lot of speculation about how a second Trump term could reshape compliance. Having worked on compliance software at both Coinbase (during Trump's presidency) and Brex (during Biden's presidency), I've been thinking about what this means for our industry. Here's my perspective on what's likely to change and how companies should prepare. ?? Three Key Changes to Watch - Crypto Regulation Shift: The biggest change might be in crypto oversight. Trump's recent pro-crypto statements and criticism of banks suggest we could see crypto regulation move from the SEC to the CFTC. This would fundamentally change how crypto companies approach compliance. - Banking Innovation: Traditional banks might find more freedom to innovate, particularly in crypto services. The rumored federal Bitcoin reserve could accelerate institutional adoption. However, banks will need robust compliance frameworks to manage the associated risks. - State vs Federal Tension: While federal regulations might relax, state-level oversight - especially in places like New York and California - will likely remain strict. Companies will need sophisticated compliance programs that can handle this regulatory patchwork. ?? The Compliance Paradox Despite Trump's deregulatory reputation, his first term actually saw record levels of OFAC enforcement actions. This highlights an interesting paradox: while broad financial regulations might ease, specific enforcement areas - especially around international money flows - could intensify. Why? Trump's America First could translate into heightened scrutiny of cross-border transactions, particularly with countries like China. Banks and fintechs might find themselves navigating fewer rules overall but facing more intense scrutiny in specific areas. ?? What This Means for Compliance Teams Having built compliance technology strategies from scratch, I think the key to navigating this changing landscape will be flexibility. Here's what compliance teams should focus on: - Build adaptable compliance frameworks that can quickly adjust to regulatory changes - Invest in technology that can handle multi-jurisdictional requirements - Maintain strong state-level compliance even if federal rules relax - Develop robust international transaction monitoring capabilities ?? Looking Ahead The reality is, we're entering uncharted territory. The next few years could see the biggest shake-up in financial compliance since Dodd-Frank. Companies that build strong, adaptable compliance programs now will be best positioned to thrive in this new environment. Want to learn more? Check out our blog post below or listen to the podcast via the link below. We discuss Trump's second term and what it means for fintechs, banking and crypto in our latest episode of Compliance Accelerated. https://lnkd.in/gjDbdGCk