Fraud Prevention

Smartnumbers Consortium

Building a cross-organisational fraud intelligence network that turned isolated data into collective protection

Opportunity

Fraud investigators were working in silos, missing critical connections that could stop criminals who hop between organisations. Compliance concerns blocked any attempt at sharing intelligence.

The Team

Product Designer, Developer, Product Manager, Compliance teams

Wins

  • Fraudster profiles grew from 250 to 961 — a 284% increase in known threats
  • 11,042 instances of structured intelligence shared across the network
  • 18,379 fraudulent phone numbers now flagged before they reach innocent victims

The Invisible Network

Criminals don't respect organisational boundaries — they exploit them. A fraudster who hits Bank A on Monday will call Bank B on Tuesday using the same phone number. But each bank only sees their piece of the puzzle.

The Outcome

A cross-organisational fraud intelligence platform that transformed how banks protect their customers. From 250 isolated profiles to 961 connected fraudsters — this is what we built.

961 Known Fraudsters
11,042 Intelligence Shares
18,379 Numbers Flagged

Scroll to find out how we did it

01

A Hunch Becomes Data

It started with a simple question: What if the same fraudster is attacking multiple organisations, and nobody knows?

💡

Hypothesis: 250 fraudster profiles across 4 customers might have crossovers that could reveal larger fraud networks.

I dug into the data. Each "fraudster profile" contained a name, a protocol for investigators, and methods of attack. But here's the catch — each organisation only saw their own view. A fraudster could have 5 different phone numbers across 4 banks, and nobody would connect the dots.

So I learned Python and wrote scripts to cross-reference these profiles. The results were startling:

Profiles Analysed 250
Customers 4
Crossover Found 52%

Breakthrough

52% of fraudster profiles had contacted more than one organisation. The fraud industry talks about collaboration, but nobody had built the bridge. The intelligence was already there — it just needed to be connected.

02

The Compliance Wall

The data was compelling. The business case was clear. But there was a massive obstacle: compliance.

I ran discovery sessions with every stakeholder. The compliance team was clear: sharing free-text data between organisations was a GDPR nightmare. One wrong field — a name, a card number, an address — and we'd be liable.

⚠️

The Blocker: The current fraudster profiles allowed free-text input. Investigators could (and did) paste PII directly into fields. We couldn't share that.

The solution? Structured data only. If we could guide users to input fraud types and methods from a controlled vocabulary, we could share intelligence without sharing personal data.

The Pivot: We shifted from "share everything" to "share structured intelligence." This became the foundation of the entire platform.

03

Designing the Network

With compliance requirements clear, I designed a system that made collaboration feel natural, not forced.

📋

Structured Intelligence

Dropdowns and controlled fields replace free text. Users pick from UK Finance standard fraud types.

🔢

One Number, One Fraudster

Phone numbers become unique identifiers. No more duplicates across organisations.

💬

Safe Communication

Built-in chat with real-time PII detection. Red highlights warn before sending.

Four Core Features

💬

Consortium Chat

Investigators message across organisations. PII detection scans every message before send.

👤

Unified Profiles

Tile-based design shows consortium-wide view. Customisable for each user's needs.

🎯

Attack Intelligence

Structured attack details from any source — profiles, calls, or chat.

🏷️

Global Names

Randomly generated consortium names replace real names. Privacy preserved.

04

Building in Slices

The scope was massive. Releasing everything at once would take months. Users were already asking for it.

We sliced the work into five incremental releases, each delivering value while building toward the full vision:

1

Foundation

Structured fraud types and methods. Users could start inputting standardised data immediately.

2

Visibility

Activity feed showing how fraudsters attack across the network. Investigators see the full picture.

3

Deep Investigation

Call history and cross-organisational timeline for each fraudster profile.

4

Collaboration

Chat launches. Investigators can now communicate directly, challenge intelligence, and share insights.

5

Automation

API updates push consortium data to downstream prevention systems. Automatic protection.

05

What We Learned

Each release taught us something unexpected.

📈

Timing Matters

First release used a button to trigger data input. Barely used. Second release triggered the modal automatically when a fraudster was assigned. 10x increase in data capture.

❤️

Demand Exceeded Supply

The feature was so anticipated that we were flooded with feature requests. Had to adopt dual-stream development to keep up.

🛡️

Trust is Everything

Users rely on Smartnumbers to be fair gatekeepers. We built a support interface for handling data challenges and disputes.

🎯

From Silos to Shield

Smartnumbers Consortium transformed isolated fraud investigations into a collective defence network. What started as a data analysis hunch became a platform that protects millions of customers.

"The best part? An investigator at Bank A now knows within seconds if a caller has already been flagged by Bank B. The fraudster's game is up."