Fraud Intelligence Report · Foresight 2027
The New Fraud Frontier
The blind spots where fraud lives today and how to get ahead of them
Key findings of the Fraud Intelligence Report 2027
The fraud that hits companies hardest arrives after onboarding and lurks in the blind spots of their systems, waiting for its moment.
01
4.5x
AI has made fraud cheaper, not reinvented it.
AI-powered fraud is now 4.5 times more profitable than traditional fraud.
INTERPOL, 2026
02
USD 3.3B
The costliest fraud happens after onboarding.
Record exposure of US lenders to synthetic identities.
TransUnion, 2025
03
94% vs 70%
Spending more isn’t working.
94% of banks invested more in 2025; 70% are still reporting losses.
Digital Trust Meet México 2026
Explore the contents
Chapter 01 · Context
How the landscape has changed
Trust used to be granted once, at onboarding, and then filed away. In 2026, fraud has learned to live in everything that happens afterwards, waiting for its moment.
- Why fraud that gets in through identity ends up booked as a credit default or credit loss
- Six key findings to understand where fraud lives today
- Continuous identity as the backbone of prevention
Key figure · 01
4.5x
More profitable: AI-powered fraud versus traditional fraud (INTERPOL, 2026)
Chapter 02 · The new frontier
AI is rewriting the economics of fraud
AI hasn’t invented a new kind of fraud: it has made it cheaper, faster and scalable. And it exploits an almost universal mistake: looking at identity only once.
- Frontier AI, within reach of defenders and attackers alike
- The KYC interception “cloak” and the verified-account factory
- The global industrialisation of fraud and fraud as a service
Key figure · 02
< 1%
Of the USD 442 billion lost to scams in 2025 was traced or recovered (GASA and INTERPOL)
Chapter 03 · The scale of the threat
Five shifts taking fraud beyond verification
Five trends are widening the threat, and they all exploit the same blind spot: verifying identity once and filing it away.
- From AI agents to synthetic identities injected at capture
- Dormant identity: invisible by design to anyone who verified it only once
- Four questions about the same identity to stay ahead
Key figure · 03
5 → 1
Five trends, one blind spot: what happens after verification
Chapter 04 · The visible fraud map
A global view, region by region
Fraud isn’t uniform: each region combines its own fraud types, regulatory maturity and new modus operandi.
- NORAM, Europe, LATAM, APAC, Africa and MENA with official 2025-2026 figures
- The dominant fraud type and the key figure for each market
- Facephi’s take on what each region means
Key figure · 04
+69%
INTERPOL fraud notices in Europe between 2024 and 2025
Chapter 05 · From compliance to proof
Seeing identity continuously
Regulation no longer asks for a one-time check: it requires knowing identity over time and justifying every decision. Those who get ahead, win.
- The three regulatory changes shaping the 2027 operating environment
- The 2027 trend in each region, from NORAM to Africa
- Know Your Agent: the next frontier of identity
Key figure · 05
July 2027
Regulation (EU) 2024/1624 extends due diligence to ongoing monitoring
Chapter 06 · Where fraud is heading
What to anticipate
Every technological advance closes some doors to fraud and opens others. What to anticipate to know where it will live tomorrow.
- The next big fraud won’t start in a bank: travel, gaming, e-commerce and the public sector
- After account takeover comes agent takeover
- Know Your Agent: what you’ll need to verify when an agent pays on your customer’s behalf
Key figure · 06
+150
Organisations backing the protocol for AI agents to identify each other (Linux Foundation, 2026)
Would you spot these five identity frauds?
Five real cases taken from the report, with no time limit: correct answers count and, as a tiebreaker, how long you take. Spoiler: in 2026, fraud is almost never where you look for it.
Before you start
Five cases, five blind spots.
How many would you have seen coming?
Five real cases taken from the report. No time limit, but the clock is running. Spoiler: in 2026, fraud is almost never where you look for it.
Map of fraud and regulation across six regions
Each region combines its own fraud types, regulatory maturity and new modus operandi. Click on the map to see what dominates in each one.
What’s inside the Fraud Intelligence Report 2027
- Regional analysis: the dominant fraud in six regions and the regulation to watch.
- What is changing in fraud with the arrival of frontier AI, from onboarding to agent takeover.
- A self-assessment: five maturity levels, ten questions for the risk committee and seven indicators.
Facephi experts on fraud, AI and agents
Expert voice · Facephi Observatory
The question is no longer just who the user is, but who, or what, is acting on their behalf.
Javier Barrachina
R&D Director, Facephi
Expert voice · Facephi Observatory
Every industry with its own data model, but with a common language to describe fraud patterns.
Mike Luparelli Mathieu
Product Innovation Director, Facephi
One platform, no blind spots for fraud
Facephi Intelligence Platform is a continuous identity platform against fraud for regulated entities. It starts from the same premise as the report: identity isn’t verified once and forgotten; it builds up and runs through the entire customer relationship.
01
Verify
Identity starts with a verification.
02
Assess risk
Every interaction adds context.
03
Decide and prove
With traceable, explainable evidence.
Trust the history behind every identity
Frequently asked questions about the Fraud Intelligence Report
It is Facephi Observatory’s annual anti-fraud intelligence report. In the Foresight 2027 edition, The New Fraud Frontier, we analyse where identity fraud lives today: after onboarding, across six regions and in the face of the arrival of AI agents.
It is an identity fabricated by combining real and fake data, which passes initial verification because it doesn’t belong to any specific victim. Its risk shows up after onboarding, once the account is already “trusted”.
It is the evolution of account takeover: the attacker no longer takes control of the customer’s account, but of the AI agent acting on their behalf. The report flags it as one of the trends to anticipate ahead of 2027.
KYC verifies who the customer is; KYA verifies whether the AI agent acting for them is authorised to take that action, at that moment. We explain it in depth in our whitepaper on Know Your Agent.