Announcing Sardine's $70M Series C - Building Trust in the Age of AI
Money moves at the speed of light. Crime moves at the speed of AI. Between these forces, trust is eroding faster than ever.
In 2023, scammers stole $1 trillion globally. That's lost savings, ruined businesses, and shattered lives. As payments accelerate and AI makes fraud more sophisticated, the gap between threat and defense keeps widening.
When we founded Sardine in 2020, we saw this coming. The existing vendor landscape was fragmented and slow-moving. Point solutions couldn't keep up with evolving threats.
As former leaders of risk and compliance teams at Coinbase, Revolut, Uber, and PayPal, we felt this pain firsthand. We built Sardine to be the platform we wished we had and knew we needed for the future.
Today, we are announcing our $70M Series C funding round led by Activant Capital, bringing our total raised to $145M. Most importantly, we’re sharing our vision for rebuilding trust in financial services through AI.
The trust gap is widening
Last year, scammers stole over $1 trillion globally. Deepfakes have made synthetic identities nearly indistinguishable from real ones. Social engineering attacks have become industrial-scale operations globally. Meanwhile, compliance costs now consume 10-20% of operational budgets for many financial services companies.
The tools to fight these threats haven't kept pace. Legacy systems generate overwhelming alert volumes. Fraud teams are drowning in manual reviews. Compliance requirements grow more complex by the day. And real-time payments mean there are only milliseconds to detect threats before money moves.
Consider a typical fraud or compliance alert. What seems simple – verifying a user's identity or validating a transaction – often requires 30+ minutes of manual investigation. Analysts piece together transaction histories, map relationship networks, verify location data, and cross-reference multiple systems. Multiply this by hundreds of alerts per day, and the backlog becomes impossible.
Building from first principles
We approached this challenge from first principles. Three key insights have guided our journey:
- All risk problems are fundamentally data problems
- The most valuable signals come from how users actually behave
- Different risk types (fraud, compliance, credit) are deeply interconnected
These insights led us to build something unique: a unified risk platform powered by the richest behavioral data in the industry and enhanced by AI agents that make human teams 10x more effective and efficient.
Momentum that matters
In the past year, we have:
- Doubled our customer base
- Grown ARR by 130%
- Profiled over 2.2 billion devices
- Expanded to 300+ enterprises across 70 countries
- Launched a suite of AI agents that automate critical risk operations
But what matters most is the impact we're driving for our customers:
- Tier-1 bank prevented 42% of wire fraud using our Sonar network
- Commercial neobank cut dispute volume by 93.75%, reducing chargeback losses by 70%
- Expense management platform reduced ACH fraud by 60%
- Global marketplace saved 15,000 hours annually in manual reviews
- Cross-border remittance provider improved fraud detection precision from 15% to 70%, reducing manual reviews by over 85%
- B2B intelligence company improved high-risk customer detection by 40x over their previous provider
The Sardine difference
Our edge comes from three core strengths:
- Richest Behavioral Intelligence: In an age of AI-generated fraud, how users behave on their devices has become the most reliable truth signal. Our Device Intelligence and Behavioral Biometrics catch what traditional systems miss. Every fraudster has a tell. And when fraudsters use deepfakes or stolen identities, their patterns give them away.
- Unified Risk Platform: We've built a complete platform for the modern risk stack. From KYC and fraud prevention to AML transaction monitoring and Credit Underwriting, everything works together. Rules and ML models can be deployed with minimal code. Risk signals are shared intelligently in real-time across modules.
- Network Effects: With over 2.2 billion devices profiled and growing, our network provides unmatched visibility into emerging threats. Our Sonar consortium enables fully-compliant data sharing across industries like the world's largest banks, e-commerce companies, marketplaces, and payment processors like FIS, Ascensus, Deel, GoDaddy, and X. Each new customer makes the network stronger. When one customer blocks a fraudulent device, the entire network is protected.
GoDaddy experienced this firsthand. As a leader in online presence and merchant payments, they needed to consolidate risk management while improving fraud detection and payment security. Sardine’s unified platform replaced multiple vendors, integrating fraud prevention, ACH risk assessment, and rule-based monitoring into a single solution. This reduced complexity and cut rule deployment time from days to hours, allowing the risk team to move faster without depending on a 16+ person engineering team.
Beyond fraud prevention, Sardine also helped GoDaddy expand into credit risk evaluation, giving them real-time insights into merchant viability during peak sales periods.
The result: fewer disputes, lower losses, and a more scalable risk infrastructure that grows with the business.
“Sardine’s platform is now the core of our risk and fraud workflows, allowing us to consolidate vendors and improve operational efficiency," said Arjun Ramakrishnan, Head of Risk for GoDaddy Payments. "By migrating to Sardine, we’ve significantly reduced our reliance on engineering, cutting rule deployment timelines from days to hours. Sardine’s flexibility has also helped us expand beyond fraud detection into credit risk evaluation to mitigate potential losses during high-sales periods.”
Introducing AI agents for risk teams
The next phase of Sardine focuses on AI agents that augment human risk teams. Today, we’re announcing four new agents. Critically: All agent decisions are fully explainable, with clear audit trails showing exactly why each choice was made.
KYC Agent
Automates the trickiest parts of onboarding – name mismatches across cultures, inconsistent date formats, and document verification challenges. Most banks and fintechs rely on documentary KYC (Drivers License, Passport, National ID card) as supplementary evidence when SSN checks fail during onboarding. Even after a document is uploaded, multiple checks are still required, such as verifying that the name on the document matches the one provided during onboarding.
Instead of forcing good customers to wait in review queues, our KYC agent resolves edge cases automatically while maintaining compliance (with human oversight).
We are achieving 88% auto resolution rates with our KYC agent thereby cutting down onboarding times and consequently also recovering potentially lost revenues.
Sanctions Screening Agent
Helps your team efficiently review sanctions, politically exposed persons (PEP), and adverse media alerts. It learns your standard operating procedures (SOP), provides detailed audit logs, and enables compliance teams to validate decisions efficiently.
While building this Agent, we have proactively engaged with regulators and examiners. And while we could fully automate the approve/decline of a sanctions alert or negative news alert. Instead, our Agent proposes a decision that a human analyst can accept/reject. This allows us to correlate and benchmark accuracy of the Agent.
To regulators, this level of transparency is critical so they can become more comfortable with AI Agents in what is the most critical part of customer due diligence. In particular, the cost of a false negative (sanctioned individual being approved) is disproportionately higher compared to false positives.
Merchant Risk Agent
A decade ago, there were thousands of sellers. Now, with innovations like “Tap to Pay,” anyone can accept card payments, and we are quickly moving toward a world with billions of sellers.
The old way of doing KYB was not built for this. PSPs and merchant acquirers designed their systems for a much smaller scale, and that model needs to evolve.
Our Merchant Risk Agent pulls data from multiple sources to determine a merchant’s “True Industry” and “True MCC” and continuously monitors for changes as well. Fraudulent merchants often register with a low-risk MCC from Visa or Mastercard, only to later switch to a high-risk category to sell restricted goods like firearms.
The Agent also summarizes complaints. Are they shipping on time? Are they issuing refunds properly? It connects this data with transaction activity to catch signs of transaction laundering and collusion fraud, such as merchants swiping thousands of stolen cards to siphon funds.
Disputes Agent
Handles the entire chargeback and dispute process, from gathering data to preparing evidence packages. It formats submissions to match processor requirements and files them automatically, reducing manual work and improving win rates.
Several of our customers have been able to automatically submit evidence packages to the card processors and schemes, cutting down time per evidence from 30+ minutes to 0. And to top it all, they see even higher chargeback win rates as there's less room for human error.
Our AI agents are in production today, delivering 4X ROI and unlocking millions in previously lost revenue.
And they're just the beginning.
Future ready
We’re always focussed on what’s coming and building that into our platform. These are our key focus areas with our Series C funding.
1. AI Agents deliver real ROI - If they comply
Banks and fintechs using our AI agents are seeing up to 4X ROI through automated onboarding and alert resolution. With full audit trails and human oversight, these agents accelerate revenue while strengthening compliance controls. Fast, fair decisions prevent customer abandonment and reduce litigation risk from unfair access denials.
2. FRAML convergence and vendor consolidation is key
Enterprises are breaking down silos between fraud prevention and AML. Our unified case management platform enables cybersecurity, fraud, and compliance teams to collaborate seamlessly. When one team spots a threat—like sophisticated bots attempting account takeovers—the entire organization benefits from that intelligence.
3. In card issuing milliseconds matter
When a customer taps their card, milliseconds matter. Our platform evaluates thousands of risk signals before the payment terminal beeps. This real-time intelligence has attracted major payment processors and fintechs to our issuing fraud solution. We give issuers unprecedented control over their risk decisions. Full ML scoring in hundreds of milliseconds (or lower in rules only mode), Asynchronous ML scoring to catch subsequent fraud, with flexible deployment options to balance speed and accuracy
4. AI vs bad AI agents
AI agents are initiating payments, booking travel and managing diaries. They're also powering sophisticated fraud at scale. Our platform detects automated threats other systems miss, from OpenAI's Operator to Selenium, Puppeteer and Playwright. We identify AI "tells" through subtle behavioral patterns: mouse movements, typing cadence, and device interactions that separate bots from humans.
5. Every digital store of value is now in scope for fraud/ML
Fraudsters are targeting any store of value, from 401(k)'s and IRA's to airline miles and gift cards. Our AI platform helps protect these account types by identifying suspicious patterns in users, behavior, devices, and data.
6. B2B payments need real-time detection
FedNow and RTP are transforming B2B transactions. With higher values and complex approval chains, business and financial institution payments need sophisticated risk controls that work in real time. Often wires have less fraud and compliance controls than consumer payments, but as the volume of fraud changes, so does the focus.
7. Customer due diligence is always on
Regulators, through FFIEC guidance, now require banks and fintechs to continually update customer risk ratings based on a complex web of signals: ownership changes, unusual transaction patterns, 314(a) law enforcement inquiries, SAR filings, and more. Our platform unifies these critical compliance functions: robust CDD, case management, transaction monitoring, and SAR filing work together as a single, intelligent system.
8. The legal landscape is shifting for payment fraud
By June 2026, Nacha will require all banks receiving ACH credits to implement fraud detection measures.This new standard of care creates both challenge and opportunity. Banks need sophisticated real-time detection for incoming payments, but building these systems internally is complex and costly. Our platform already provides the capabilities these institutions need: automated name matching, account age analysis, and behavioral pattern detection that works at ACH speed and scale.
Just getting started
This funding will accelerate our mission to rebuild trust in financial services. We're expanding our enterprise capabilities, growing globally, and advancing our AI agent platform.
The challenges we face are immense. Financial crime is more sophisticated than ever. But we've assembled a team of risk nerds, data scientists, and engineers who live for these problems. We're building the tools we wish we had. And we're just getting started.
For our customers, investors, and team who believe in this mission – thank you. The real work lies ahead.
Want to join us? We're hiring across engineering, data science, sales, and customer success.
Learn more at sardine.ai/careers
.png)