How HSBC uses AI

Mastercard completed its first live, authenticated agentic transaction in Hong Kong with HSBC and DBS Hong Kong as issuing banks. An AI agent from CardInfoLink booked a ride to Hong Kong International Airport through the mobility provider hoppa and paid for it with a tokenized credential authenticated with Mastercard Payment Passkeys. Each agent receives its own Mastercard Agentic Token, and the consumer's consent is captured explicitly.

Mastercard says similar transactions were supported on all Mastercard card types issued by Hong Kong banks, including Citi Hong Kong, Hang Seng Bank, Standard Chartered Hong Kong and Mox Bank. HSBC says it is enabling its credit and debit card customers in Hong Kong to pay this way, with clear consent and transparency. It is a milestone transaction, and no outcome figures are disclosed.

Google Cloud reports that HSBC needed compliant, multimarket and multilingual marketing, and that working with Capgemini it extended its existing copy generation tool, ComAmplifier, with a coordinated set of agents that collaborate in a single workflow to produce targeted content and messaging, including insight, copy, imagery and approvals. The entry gives no metric and does not say how the compliance step works.

HSBC has deployed a generative AI assistant to its Corporate and Institutional Banking servicing teams that supports 3 million client interactions a year, which HSBC says reduces turnaround times and improves the client experience. The same sentence says 88% of clients rate HSBC easy to deal with, a bank wide score that the page does not attribute to the assistant, so it is not recorded as a metric.

HSBC built Smart Checking in house, an AI system that ingests trade documents, extracts and classifies the data, interprets the conditions of each documentary credit and runs the checks. Low confidence results and complex cases go to trade specialists. HSBC says it processes more than a million documentary presentations a year and that Smart Checking is used in Hong Kong, the UAE and the UK, with a global rollout to follow. No result figures are published.

HSBC Private Bank launched Wealth Intelligence, a generative AI powered platform built by its own in house developers and powered by OpenAI's large language model, for its client facing staff: investment counsellors and product specialists. It gives them fast, simple access to the full range of reports published by the bank's Chief Investment Office, so they can bring quality market insights and personalized investment strategies into client conversations. HSBC says the next stage will let Wealth Intelligence screen and identify suitable products for the bank's wealth management teams to discuss investment options and asset allocation with clients. It launched first in Hong Kong and Singapore, with HSBC planning to scale it to more markets.

Silent Eight has supplied HSBC with automation for name screening and adverse media alerts, and in February 2024 the two expanded the partnership to automated alert closure for transaction screening, which investigates and resolves payment screening alerts in real time. No outcome figures are disclosed.

HSBC replaced rule based transaction monitoring with Dynamic Risk Assessment in its key markets, a machine learning system co developed with Google Cloud that scores customers for money laundering risk from their full transaction and customer data, and routes the highest risk to investigators. HSBC piloted it in 2021 and says it checks about 980 million transactions a month for signs of financial crime. HSBC reports finding two to four times more financial crime with much greater accuracy, and cutting the processing time to analyze billions of transactions from several weeks to a few days.

HSBC Hong Kong built a Budget feature into the HSBC HK app using AI and analytics on more than 500 million transactions its retail customers carried out in 2021, worth about HK$3 trillion. The feature shows customers how they spend and gives insights tailored to their circumstances and goals across cash flow, budgeting, expenses, investment and protection, so customers can apply money management techniques even if they usually struggle with budgeting.

In September 2019 HSBC's Global Trade and Receivables Finance business announced two systems for financial crime detection. One is an anti money laundering surveillance system built with Quantexa that uses network analytics and contextual monitoring to screen trade finance transactions against more than 50 money laundering scenarios; it was live in the UK and Hong Kong.

The other is a first line sanctions check that HSBC built in house with advanced algorithms and machine learning, live in India with a rollout to 41 markets planned by the end of 2019. The release gives the bank's overall volumes (more than 5.8 million trade transactions screened a year, about one million sanctions submissions a month) but no figure for what either system changed.