---
title: 'HSBC: Dynamic Risk Assessment transaction monitoring'
slug: hsbc-dynamic-risk-assessment
stable_id: 953d873c30c6de37
company: HSBC
function_code: financial_crime
pattern_codes:
  - exception_based_operations
  - creator_to_judge
evidence_strength: mixed
publication_tier: showcase
freshness: current
reviewed_at: '2026-08-23'
updated_at: '2026-08-22'
source_quality_summary: 2 primary, 1 vendor; publication outcomes are verified and reported.
caveat_summary: >-
  Company-reported comparison without independent outcome validation; 'suspicious
  activity' and 'financial crime found' are not defined as confirmed offenses.
collections:
  - regulated-autonomy
bundle_version: 1.0.0
bundle_fingerprint: sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93
canonical_url: https://brianletort.ai/transformations/hsbc-dynamic-risk-assessment
---

# HSBC: Dynamic Risk Assessment transaction monitoring

Transaction-monitoring teams do not need a large rules-generated queue to preserve risk coverage; machine-learning risk detection can narrow the queue while investigators retain consequential review.

Function: Financial crime. Patterns: Exception-based operations; Creator to judge. Evidence: mixed.

Freshness: current. Reviewed: 2026-08-23. Updated: 2026-08-22.


Source quality: 2 primary, 1 vendor; publication outcomes are verified and reported.

## Before

1. **Rules-based transaction-monitoring system** — Screen transactions against predefined parameters and generate alerts. (control: Rules encode the monitoring thresholds.)
2. **Financial-crime investigators** — Manually review flagged transactions, including a high volume of false positives. (control: Investigators determine whether flagged activity warrants action.)

## After

1. **Dynamic Risk Assessment** — Score transactions and identify suspicious activity using an AI model trained on HSBC data. (control: HSBC trains and assesses the model under its responsible-AI practices.)
2. **Financial-crime investigators** — Review a smaller, more risk-concentrated alert queue. (control: Humans retain investigation and escalation decisions.)

## Decision rights

The AI system prioritizes risk; HSBC investigators retain case review, customer-contact, escalation, and reporting decisions.

## Exception path

Suspicious or uncertain activity is routed to investigators for manual review and escalation.

## Outcomes

- **Transaction-monitoring alert volume** (reported): Rules-based monitoring during the 12 months before go-live. → HSBC reported 60% fewer alerts while identifying two to four times as much suspicious activity.. Company-reported comparison without independent outcome validation; 'suspicious activity' and 'financial crime found' are not defined as confirmed offenses.

## Executive lesson

The defensible value is not automation of the compliance decision; it is a smaller, higher-yield investigation queue with humans retaining consequential case decisions.

## Anti-pattern

Treating alert reduction as proof of crime prevention or allowing the model to close consequential cases without investigator review.

## Questions for leaders

- Which queue metric will be compared before and after deployment: alerts, cases, or confirmed suspicious activity?
- Which decisions remain with investigators?
- How will model drift and missed-risk rates be monitored?

## Sources

- [Harnessing the power of AI to fight financial crime](https://www.hsbc.com/news-and-views/views/hsbc-views/harnessing-the-power-of-ai-to-fight-financial-crime) — HSBC
- [HSBC Holdings plc Annual Report and Accounts 2023: Risk review](https://www.hsbc.com/-/files/hsbc/investors/hsbc-results/2023/annual/pdfs/hsbc-holdings-plc/240221-risk-review-2023-ara.pdf) — HSBC
- [Fighting money launderers with artificial intelligence at HSBC](https://cloud.google.com/blog/topics/financial-services/how-hsbc-fights-money-launderers-with-artificial-intelligence) — Google Cloud, authored by HSBC