The blockchain analytics firm that governments and exchanges rely on to trace crypto crime is now deploying AI agents to do the investigating. Chainalysis announced blockchain intelligence agents at its Links 2026 conference in New York — trained on more than 10 million investigations spanning a decade of on-chain analysis.

The timing is deliberate. Chainalysis’s own data shows AI-powered crypto scams are exploding: $17 billion stolen in 2025, AI-enabled fraud 4.5 times more profitable than traditional methods, deepfake-driven scams hitting $4.6 billion. The company is deploying AI to fight AI.

This analysis draws on Chainalysis’s official blog post, the Links 2026 recap, CoinDesk, CoinGeek, PYMNTS, Chainalysis’s 2026 Crypto Crime Report, and FBI Internet Crime Report data via CoinDesk — we research and analyze rather than testing products hands-on. Rob Nugen operates ChatForest; the site’s content is researched and written by AI.

The Crime Problem Driving the Investment

Before the agents, the numbers.

Chainalysis’s 2026 Crypto Crime Report and the FBI’s Internet Crime Report paint a picture of fraud scaling faster than human investigators can keep up:

Metric Number Source
Total crypto stolen via scams/fraud (2025) $17 billion Chainalysis
US crypto scam losses (2025) $11.4 billion FBI
Year-over-year increase (US losses) 22% FBI
AI-enabled scam profitability vs. traditional 4.5x Chainalysis
Deepfake-driven scam losses $4.6 billion Chainalysis
Deepfakes as share of high-value cases 40% Chainalysis
Impersonation scam growth (YoY) 1,400% Chainalysis

Three dynamics are converging:

AI makes fraud cheaper and more convincing. Large language models enable scams to cross language barriers. AI-generated deepfakes, voice clones, and synthetic personas reduce the cost of creating believable identities. A scammer who previously needed native-language fluency and video production skills now needs a laptop and an API key.

Impersonation is the dominant vector. The 1,400% year-over-year growth in impersonation scams is staggering. Criminals pose as legitimate organizations, government officials, or investment advisors using AI-generated content that is increasingly difficult to distinguish from real communications.

Professional criminal infrastructure is emerging. TRM Labs documented $23 billion in verified fraud plus an additional $12 billion tied to community complaints. Chinese money laundering networks have emerged as a dominant force, offering laundering-as-a-service and other specialized criminal infrastructure. This is organized crime operating at industrial scale.

The result: investigation teams are overwhelmed. Every analyst hour spent on a false positive is an hour not spent on a real case. This is the bottleneck Chainalysis’s agents are designed to break.

What the Blockchain Intelligence Agents Do

The agents are built into Chainalysis’s existing platform, not a separate product. They operate through natural language — investigators describe what they need in plain English rather than learning specialized query syntax.

Six Core Capabilities

1. Alert Enrichment and Automation

Raw compliance alerts are the starting point for most investigations. Currently, analysts manually pull context from across the platform for each alert. The agents automate this: take a raw alert, pull context, enrich it with on-chain and off-chain data, and when appropriate, automatically dismiss or escalate it.

This addresses the highest-volume bottleneck in crypto compliance. Exchanges and financial institutions process thousands of alerts daily, most of which are false positives. Automating triage lets human analysts focus on cases that actually require judgment.

2. Summary Reports

The agents generate structured, defensible intelligence reports — the kind that would take analysts hours to compile manually. “Defensible” is the key word: these reports need to hold up in legal proceedings, regulatory examinations, and law enforcement referrals. The agents are designed to produce audit-ready documentation, not just summaries.

3. Custom Web Applications

Investigators can use the agents to build custom tools and dashboards for specific investigative or compliance needs. This turns ad-hoc analysis into repeatable workflows without requiring engineering support.

4. Transaction Analysis

Finding and flagging activity across specific time windows with precision, at scale. This is the core analytical work — tracing funds through mixers, bridges, and layered transactions — accelerated by agents that can process patterns faster than human analysts.

5. Intelligence Collection

The agents augment investigations with open-source intelligence (OSINT), gathered and organized automatically. In blockchain investigations, OSINT often connects on-chain addresses to real-world identities through social media posts, forum activity, domain registrations, and other public data.

6. Team Orchestration

Set up agents to monitor on-chain activity, alert on suspicious patterns, conduct automated analysis, and surface leads to humans for action. This creates a continuous monitoring layer that doesn’t require analysts to be actively watching dashboards.

The Training Data Advantage

The agents are trained on a proprietary dataset built over more than a decade: billions of screened transactions and more than 10 million investigations conducted inside Chainalysis’s Reactor platform.

This is Chainalysis’s most defensible competitive advantage. The dataset represents:

  • Real investigation workflows — not synthetic training data, but actual patterns of how experienced investigators trace funds, identify clusters, and build cases
  • Outcome data — which investigative approaches led to successful asset recoveries, prosecutions, and compliance actions
  • Cross-jurisdictional patterns — cases spanning multiple countries, blockchains, and regulatory frameworks

Chainalysis engineered the agents around four design principles:

  1. Data quality — the training data comes from vetted, production investigations, not web scrapes or synthetic datasets
  2. Context and reasoning — agents are designed to explain their analytical steps, not just produce outputs
  3. Auditable results — every agent action is traceable, critical for legal and regulatory use
  4. Deterministic workflows — the same inputs, rules, and data produce the same outcomes, ensuring reproducibility

The auditability and determinism requirements are notable. In law enforcement and compliance, an AI system that produces different results on the same inputs is useless — or worse, a liability. Chainalysis is explicitly designing against the “creative” hallucination tendencies of general-purpose LLMs.

The agent announcement was the headline at Links 2026 (March 31-April 1, New York Marriott Marquis), but the broader conference reinforced three themes:

AI is amplifying both sides. Every panel acknowledged that the same AI capabilities making investigations more efficient are also making fraud more sophisticated. This isn’t a one-time advantage — it’s an ongoing arms race.

Crypto and traditional finance have converged. The distinction between “crypto crime” and “financial crime” is dissolving. Criminal networks use crypto for laundering proceeds from fraud that originates in traditional finance, and vice versa. Investigation tools need to span both worlds.

Collaboration scales intelligence. The Microsoft Digital Crimes Unit presented a case study on Operation Trashpanda, using Chainalysis Reactor to trace payments to Nigerian exchanges and working with the US Secret Service and Nigerian Police to seize 338 malicious domains. UK prosecutors detailed “Operation Man,” using Chainalysis to trace $47 million in bitcoin and seize $4.25 billion from a Chinese Ponzi scheme. These cases illustrate that effective crypto crime investigation requires public-private-international coordination — exactly the kind of complex workflow that AI agents could help orchestrate.

The conference awarded its Public-Private Partnership Award to the UAE Cybersecurity Council and its Innovation Award to MoonPay.

Competitive Landscape

Chainalysis dominates blockchain analytics but faces growing competition in the AI agent space specifically.

The Key Players

Company AI Capability Training Data Market Position
Chainalysis Blockchain intelligence agents (summer 2026) 10M+ investigations, decade of Reactor data Market leader, $8.6B peak valuation
TRM Labs AI agent for natural language on-chain analysis Verified fraud data ($23B+ in 2025) Strong second, growing government contracts
Elliptic AI-powered “copilot” for compliance Deep learning for laundering pattern detection UK-based, compliance-focused
CipherTrace (Mastercard) Integrated fraud detection Mastercard transaction data Acquired, integrated into payment network

Chainalysis’s Position

Chainalysis’s peak valuation of $8.6 billion (2022) has likely compressed — secondary market activity in 2024 suggested around $2.5 billion. The AI agent launch is partly a bid to justify premium pricing and re-accelerate growth.

The company’s advantage is the training data moat. Ten million investigations across a decade of blockchain evolution — including the rise and fall of major exchanges, the DeFi explosion, the NFT bubble, and multiple market cycles — represents institutional knowledge that competitors cannot easily replicate.

The risk: if general-purpose AI agents become capable enough at on-chain analysis, Chainalysis’s proprietary dataset becomes less decisive. A sufficiently capable LLM with access to public blockchain data and OSINT tools could approximate many investigative workflows. Chainalysis is betting that the specialized knowledge embedded in 10 million investigations provides a durable edge over general-purpose AI.

The Broader AI-in-Blockchain-Analytics Market

The AI in blockchain analytics market is expected to grow from approximately $657 million in 2025 to $3.46 billion by 2034, a 27.1% compound annual growth rate. This growth reflects both the increasing sophistication of crypto crime and the expanding regulatory requirements for crypto compliance globally.

The AI-vs-AI Arms Race

This is the most significant dynamic in the announcement. Chainalysis is explicitly framing its agents as a response to AI-enabled crime:

Criminals use AI to scale fraud. LLMs generate convincing phishing messages in any language. Deepfakes create synthetic identities for impersonation scams. AI automates the creation of fake investment platforms. The result: fraud operations that previously required dozens of people can now be run by a handful with AI tools.

Investigators use AI to scale analysis. Chainalysis’s agents automate alert triage, report generation, and pattern detection. They process transaction chains faster than human analysts and can monitor continuously without fatigue.

The asymmetry problem. Criminals need to succeed once. Investigators need to trace every transaction in a chain. AI agents help reduce this asymmetry by making the investigator’s side of the equation dramatically faster — but they don’t eliminate it. A sophisticated criminal operation using AI to generate novel laundering patterns will still challenge AI-powered detection, especially if the criminal understands how the detection systems work.

The data quality question. Criminal AI operates on public data and social engineering templates. Chainalysis’s AI operates on proprietary investigation data. This gives the investigative side a structural advantage — for now. If criminal organizations begin systematically studying how blockchain analytics firms trace funds (which sophisticated groups already do), the investigative advantage could narrow.

Rollout and Availability

Chainalysis plans a phased rollout beginning summer 2026, initially targeting:

  1. Investigations — agents that assist with fund tracing, entity identification, and case building
  2. Compliance — agents that automate alert triage, suspicious activity reporting, and regulatory workflows

The agents will be available to existing Chainalysis platform customers. Pricing has not been announced, but given Chainalysis’s enterprise positioning, expect the agents to be bundled with or added to existing Reactor and compliance product subscriptions.

No timeline has been given for broader availability or for additional agent capabilities beyond the initial six.

What This Means

For crypto compliance teams: This is the most significant tooling upgrade in years. Automated alert triage alone could free 40-60% of analyst time currently spent on false positives. The question is execution — whether the agents produce reliably defensible outputs from day one, or require a training period where humans still review every agent decision.

For law enforcement: AI agents that can generate structured reports and trace funds at machine speed could dramatically accelerate case building. The public-private partnership cases showcased at Links 2026 suggest Chainalysis is already working with agencies on agent-assisted investigations.

For the broader AI agent ecosystem: This is a concrete example of AI agents being deployed for high-stakes, adversarial work — not just productivity features or customer service. The design principles (auditability, determinism, defensible outputs) represent a template for how AI agents should be built for regulated, legal, and security applications.

For criminals: The window where AI gave fraud a one-sided advantage is closing. Not because AI agents will catch everything — they won’t — but because the cost-benefit calculation shifts when investigation can scale at the same speed as the fraud.