Quantifind: Interview With Founder And CEO Ari Tuchman About AI-Powered Risk Intelligence, Financial Crime Prevention, And Agentic AI

Quantifind is an AI-powered risk intelligence company that helps financial institutions, government agencies, and enterprises detect financial crime, identify sanctions exposure, investigate complex networks, and manage regulatory compliance. Pulse 2.0 interviewed Quantifind Founder and CEO Ari Tuchman to learn more about his background, the company’s AI-native risk intelligence platform, emerging financial crime threats, responsible agentic AI, its recent $200 million investment, and the future of risk management.

Ari Tuchman’s Background

When asked about his professional background and the experiences that led him to establish Quantifind, Tuchman shared:

My career began in academic and experimental physics, focusing on signal processing and precision measurement—including hazardous material detection and quantum sensors. I founded Quantifind with the core belief that the signal extraction techniques used to isolate signals in quantum measurement could be more broadly applicable. Today, as CEO, I lead our team in building AI-powered SaaS solutions that transform vast, unstructured open-source data into actionable risk intelligence for global institutions.

Quantifind’s Core Solutions

Discussing Quantifind’s technology and its applications in financial crime prevention, compliance, and national security, Tuchman explained:

Trusted by 10 of the top 20 global financial institutions, Quantifind delivers AI-native Risk Intelligence that transforms fragmented data into actionable insights, helping organizations detect threats others miss, dramatically reduce false positives, and make critical decisions with confidence.

Powered by Graphyte, Quantifind’s AI-native Risk Intelligence platform, our purpose-built solutions address some of the most complex challenges in financial crime prevention, compliance, and national security:

Know Your Customer (KYC): Accelerate onboarding, customer due diligence, and ongoing monitoring with more accurate identity resolution, adverse media screening, and relationship intelligence. Uncover hidden customer risks while reducing false positives and manual reviews.

Sanctions Compliance: Detect direct and indirect sanctions exposure across individuals, businesses, and complex networks. Improve screening accuracy, identify hidden ownership and evasion risks, and reduce the volume of unnecessary alerts.

Transaction Intelligence: Go beyond traditional transaction monitoring by enriching payments and financial activity with external risk intelligence. Identify suspicious counterparties, hidden relationships, and emerging risks that transaction data alone cannot reveal.

Investigations: Turn complex, time-consuming research into faster, evidence-backed investigations. Connect individuals, entities, and networks across billions of data points to uncover financial crime risks and accelerate case resolution.

Third-Party Risk Management: Expose hidden ownership, sanctions exposure, foreign influence, and reputational risks across vendors, suppliers, partners, and broader business networks.

Graphyte combines purpose-built AI with entity resolution, and billions of data points spanning global news, corporate registries, sanctions and watchlists, court records, and other public and commercial sources. It reveals connections traditional screening and monitoring tools often miss, delivering explainable findings backed by evidence.

Through Agentic Middleware and Agentic Execution, Quantifind extends this intelligence directly into enterprise systems, analyst workflows, and AI agents. Organizations can automate risk enrichment, investigation, triage, and case resolution while maintaining governance, transparency, and human oversight.

The impact is measurable: fewer false positives, faster investigations, greater analyst productivity, and the ability to analyze millions of alerts at enterprise scale. Quantifind helps organizations move beyond managing alerts to identifying the risks that matter and acting on them with confidence.

Delivered as a secure, scalable SaaS platform, Graphyte integrates seamlessly into existing technology environments, helping organizations modernize risk operations without replacing their core systems.

How Sophisticated Actors Evade Sanctions Screening

When asked how sophisticated actors use shell companies, intermediaries, and complex ownership structures to evade sanctions, Tuchman described the importance of understanding relationships rather than relying solely on name matching:

They operate through networks. They use shell companies, intermediaries, family members, ownership changes, and identity variations to create distance between themselves and the activity they are trying to conduct.

That means screening cannot stop with asking whether a name matches a sanctioned entity. We need to understand who that entity is, who owns or controls it, which people and organizations it’s connected to, and how those relationships change over time. The shift from matching names to resolving identities and understanding networks is becoming essential to effective sanctions compliance.

The Importance Of Identity Resolution

Discussing how alternate spellings, aliases, transliterations, and layered ownership structures complicate risk detection, Tuchman emphasized the importance of accurate identity resolution:

Before you can evaluate risk, you must know who or what you are evaluating. If records that belong together remain separated, the system can miss the pattern. If unrelated records are combined, it can create false suspicion and unnecessary work.

Good resolution therefore requires more than fuzzy name matching. It uses contextual evidence to test competing hypotheses: Are these records the same entity? Are they different entities with similar names? Is this ownership link current? Does the geography, chronology, or network support the connection?

This is one reason the problem aligns so closely with my work in precision measurement. The most useful answer is not about asserting confidence from limited evidence. It is about extracting a reliable signal, quantifying uncertainty, and knowing what additional evidence would change the conclusion.

Weaknesses In Traditional Sanctions Screening

Addressing the limitations of conventional sanctions screening systems, Tuchman highlighted the challenges of excessive false positives and sophisticated actors who evade straightforward matching methods:

Traditional screening has often been optimized to find possible matches rather than determine whether those matches represent actual risk. That creates two problems: large numbers of false positives that consume investigators’ time, and the risk of missing sophisticated actors who don’t appear as obvious matches.

Identity resolution sits at the center of solving both. A system must distinguish between two people who look similar but are not the same person, while also recognizing when two very different-looking records refer to the same person or organization. Names matter, but so do transliterations, aliases, locations, ownership, associates, corporate records, watchlists, and other contextual evidence.

The goal is not to relax controls. It is to make them more discriminating—reducing avoidable work while improving the chance that genuinely important risk reaches an investigator.

Identifying Hidden Supply Chain Risks

Explaining why traditional vendor-vetting approaches struggle to detect supply chain risks involving indirect relationships, Tuchman discussed the limitations of evaluating individual companies without examining their broader networks:

The organization an agency contracts with may not be where the actual risk resides. Exposure can sit several relationships away—in an owner, subsidiary, supplier, executive, foreign affiliate, or another entity embedded in the network. Reviewing only the company directly in front of you can create a false sense of security.

Traditional vetting is often point-in-time and record-by-record. Supply chains are dynamic, and the relevant information is fragmented across corporate records, sanctions and watchlists, adverse media, enforcement data, and other public or commercial sources. The challenge is to resolve those records into real entities and connect the relationships among them.

For governments, particularly where mission assurance or national security is involved, this is fundamentally a network-intelligence problem. The aim is not to label every indirect connection as risk. It is to identify which connections are material, explain why, and support a proportionate response.

Responsible Agentic AI In Regulated Environments

When asked what responsible agentic AI looks like in financial crime prevention and other regulated risk environments, Tuchman emphasized explainability, evidence-based decisions, and governance:

An agent can gather evidence, reconcile identities, investigate relationships, reason across multiple sources, and execute defined parts of a workflow. In a regulated environment, however, capability must be paired with governance.

A responsible agent operates within explicit permissions and objectives. It uses trusted sources, records its steps, shows the evidence behind its conclusions, applies confidence thresholds, and escalates exceptions. Its work should be reproducible enough for an investigator, auditor, or regulator to understand how it reached the result.

The objective is not autonomy for its own sake. It is to increase the amount and quality of analysis we can perform while preserving accountability, explainability, and control.

Balancing AI Autonomy With Human Oversight

Discussing where organizations can safely deploy autonomous AI agents and where human decision-making remains essential, Tuchman outlined how responsibilities should be divided between technology and experienced investigators:

AI is well suited to high-volume, evidence-intensive work: collecting information, normalizing records, resolving entities, mapping relationships, checking defined sources, and preparing a case narrative with citations. Those activities are time-consuming, but they can be bounded and tested.

Humans remain essential when a decision carries significant legal, regulatory, financial, or national-security consequences; when evidence is conflicting; when policy requires discretion; or when the cost of a mistake is unusually high. A sanctions disposition, an adverse action, or a national-security judgment should not be reduced to an unexplained machine output.

The practical question is not “human or AI.” It is how to divide the work intelligently. Let machines operate at a scale people cannot and give people better evidence on which to exercise judgment.

Why Better Risk Intelligence Matters More Than Larger AI Models

Addressing whether competitive advantage in AI-powered risk management will come from larger models or better underlying intelligence, Tuchman emphasized the importance of reliable data, identity resolution, and domain-specific context:

The biggest model does not necessarily win in risk.

A general-purpose model can help reason over information, but the quality of the result depends on the intelligence underneath it: reliable data, precise entity resolution, domain-specific context, source provenance, and controls that make the reasoning inspectable.

The strongest systems will combine capable models with a purpose-built intelligence layer. That combination is more important than model size alone because risk decisions have to be supported by evidence, not simply expressed persuasively.

$200 Million Investment And International Expansion

Discussing Quantifind’s recent financing and growth plans, Tuchman highlighted the company’s $200 million investment and its intention to expand its international presence:

We recently received a $200 million investment, led by Summit Partners with participation from investors like Citi Ventures, S&P Global, Deloitte and more. It’s an incredible vote of confidence in Quantifind’s missions and technology.

The funding is being used to expand Quantifind internationally across Europe, Asia-Pacific and the Americas as we partner with various organizations and agencies to advance regulatory alignment and strengthen localized risk capabilities to combat increasingly sophisticated financial crime.

The Future Of AI-Powered Risk Operations

Looking ahead over the next three to five years, Tuchman discussed how organizations can move beyond adding AI tools to fundamentally redesigning their risk management operations:

The successful organizations will redesign how risk decisions are made. Today, analysts and investigators often move among systems, search multiple sources, reconcile identities, and manually assemble the context required for a decision. Adding an AI interface without changing that underlying process will have limited impact.

I expect intelligence to be continuously assembled around entities and networks. Governed agents—whether developed and operated by organizations using our trusted intelligence through what we call Agentic Middleware, or built by Quantifind for our clients—will perform substantial portions of evidence gathering, resolution, monitoring, and case preparation. Humans will focus on exceptions, policy judgments, and decisions where accountability and experience matter most.

The differentiators will be disciplined implementation: clear objectives, high-quality data, measurable accuracy, interoperability with existing workflows, transparent governance, and feedback from expert users. The organizations that treat AI as an operating-model change—not a feature—will be in the best position to improve both effectiveness and efficiency.