From the team

Fraud Intelligence.

Research, intercept analysis, and practitioner notes from the AVIEL team. We write about APP fraud mechanics, PSP liability under FCA Consumer Duty, RIPA 2000 interception law, and the technical gap between transaction monitoring and the conversation layer where scams happen.

Abstract concept of adaptive AI-assisted fraud and behavioural-intent modelling
8 min read

Real-Time Scammer Adaptation in the LLM Era: What Fraud Teams Need to Know

Scammers using LLM-generated scripts adapt faster to intervention signals than human-scripted operators. The detection challenge has shifted from pattern recognition to behavioural-intent modelling.

Abstract regulatory framework and financial data concept for PSR reimbursement analysis
11 min read

PSR Mandatory Reimbursement at Six Months: What the Data Shows

The Payment Systems Regulator mandatory reimbursement rule took effect October 2025. Six months of operational data reveals who's absorbing the losses — and which PSPs saw fraud rates actually decline.

Abstract technical architecture diagram representing honeybot deployment patterns in payment systems
10 min read

Honeybot Deployment Patterns for Enterprise PSPs: What We've Learned

After twelve months in production with three mid-market PSPs, here's what we know about honeybot placement, conversation latency budgets, and false-positive suppression at scale.

Abstract concept of emotional manipulation and payment interception in romance fraud
9 min read

Romance Scams Hit £300M in APP Losses — Where Payment Interception Makes the Difference

Romance-driven APP fraud now ranks second behind investment scams by total value. The conversation window is longer — which means the interception opportunity is larger too.

Abstract open banking and payment authentication concept with fraud interception
8 min read

Open Banking, PSD2, and the APP Fraud Gap PSPs Haven't Closed

PSD2 strong customer authentication reduced card fraud. It did nothing for APP fraud — because the customer authorises the transfer. The gap is in the conversation, not the credential.

Abstract AI system operating in fraud operations context
11 min read

AI in Fraud Ops: What's Hype and What Actually Works in 2025

ML models for transaction scoring? Yes. Generative AI for scammer conversation simulation? Harder than vendors admit. Here's our honest take from the field.

Abstract UK GDPR and data sharing concept for payment service providers
10 min read

Sharing Fraud Intelligence Between PSPs: UK GDPR Constraints and Paths Forward

Can two PSPs share scammer fingerprint data to prevent repeat-attempt fraud? Yes — with the right DPA, consent model, and data minimisation approach.

Abstract concept of chargeback reduction through proactive fraud interception
7 min read

How Digital Banks Are Reducing APP Chargeback Exposure in 2025

The shift from post-transfer recovery to pre-transfer interception is happening now. What the forward-thinking digital banks are doing differently.

Abstract concept representing legal framework and interception in UK law
12 min read

RIPA 2000 and PSP Interception: A Plain-English Guide for Fraud Teams

Honeybots intercept conversations. Is that legal under UK law? Yes — at the PSP's instruction and under specific RIPA 2000 provisions. Here's the framework.

Abstract concept showing the conversation layer between fraudster and victim in payment systems
9 min read

The Conversation Layer: The Fraud Signal Your Stack Is Missing

Transaction monitoring sees numbers. Behavioral analytics sees patterns. Neither sees the conversation where the victim is being coached to make the transfer.

Abstract rising data trend concept representing investment fraud surge in UK
8 min read

UK Investment Scam Volumes Doubled in 18 Months — What PSPs Are Facing

Action Fraud reports and UK Finance data show investment scam APP transfers now account for 38% of total APP losses by value. Here's the pattern.

Abstract digital fingerprint concept representing behavioral identity profiling
10 min read

What Does a Scammer Fingerprint Actually Contain?

Device ID, phone number metadata, message cadence, linguistic pattern, time-zone behavior. A scammer fingerprint is a behavioral signature — not a name.

Abstract technical integration concept for payment system APIs
7 min read

Three API Patterns PSPs Use to Wire In Fraud Detection

Webhook listeners, notification bus injection, or sidecar API — how PSPs integrate real-time fraud signal tools without rerouting the payment flow.

Abstract concept contrasting static trap versus dynamic agent in fraud detection
6 min read

Honeypots vs Honeybots: Why Static Traps Miss APP Fraud

Traditional honeypots wait passively. A honeybot enters the live conversation — matching the scammer's context, extracting a fingerprint, stalling the transfer.

Abstract concept representing social engineering tactics in payment systems
8 min read

How Social Engineering Moves Through Payment Rails — And Where It Stalls

The scammer's kill chain ends at the confirmation click. Understanding the conversation layer between social engineering and fund movement is the key to pre-transfer interception.

Abstract regulatory compliance concept for UK financial services
11 min read

FCA Consumer Duty and APP Reimbursement: What PSPs Must Know by 2025

The mandatory APP reimbursement rules coming into force mean PSPs absorb losses they previously disputed. The intervention imperative has never been clearer.

Conceptual dark image representing the mechanics of authorised push payment fraud
9 min read

What Is Authorised Push Payment Fraud? A Guide for PSP Teams

APP fraud is different from card fraud and unauthorised account access. This guide explains the mechanics — and why traditional fraud tools miss it entirely.

Abstract visualization of financial data flows representing APP fraud losses in UK
7 min read

APP Fraud Is Now a PSP Liability Problem — Here's the 2024 Picture

UK Finance data shows APP losses hit £460M in 2023. Under FCA Consumer Duty, PSPs are no longer just the rails — they're co-responsible for the outcome.