Generative AI at the front, agents in the back office: Trust runs a GenAI chatbot and conversational card-spend search for customers, and Claude-based agents on Amazon Bedrock for incident triage, on a cloud stack built entirely on AWS.
How is Trust Bank using AI?
Trust Bank, the Standard Chartered and FairPrice Group bank in Singapore, uses generative AI for customer service: its Trust AI chatbot handles nearly half of customer chats end to end and halved chats needing human agents, and Trust AI Ask (September 2026) answers questions about card spending in plain language. In operations, Claude-based agents on Amazon Bedrock cut incident triage from 15–20 minutes to about two. Trust, with over 1 million customers, turned profitable in March 2026 after a S$54 million loss in 2025.
Category
Neobank
Founded
2020
Headquarters
Singapore, Singapore
Licence and supervision
Singapore full bank licence (Trust Bank Singapore Limited, formerly SC Bank Solutions), granted by the Monetary Authority of Singapore on 30 December 2020 under Standard Chartered's enhanced Significantly Rooted Foreign Bank privileges; it is not one of MAS's digital full bank licences. Supervised by MAS.
Scale
More than 1 million customers in Singapore, as of 31 July 2026 (Standard Chartered and Trust Bank)
Flagship platform
Trust AI (GenAI chatbot) and Trust AI Ask
AI leadership
Dwaipayan Sadhu · Srinivas Patil · Naveen Sethia · Angela Yeo · Aditya Gupta
- Trust says it was the first bank in Singapore to run a fully integrated generative AI customer-service chatbot with real-time handover to human agents; it handles nearly 50% of chats end to end, and complaints fell about 40% within months of rollout.
- On 8 September 2026 Trust launched Trust AI Ask, which answers natural-language questions about card spending and calculates totals; Trust plans to extend it to accounts, transfers and rewards.
- Trust runs AI agents on Amazon Bedrock AgentCore that pull logs and post a likely root cause before engineers join an incident call, cutting triage from 15–20 minutes to about two; about 65% of the analyses are actionable, and the agents recommend rather than act.
- CTO Srinivas Patil told Computer Weekly that Trust uses Anthropic's Claude Opus 5 for reasoning and Claude Haiku for lower-order tasks, with most of its models from Anthropic, while experimenting with OpenAI and open-source models.
- Trust reached profitability in March 2026; it lost S$54 million in 2025 (2024: S$93 million) on revenue of S$135 million, and says 2025 revenue grew 39% while costs fell 7%, which it credits partly to automation and early AI adoption.
- Trust holds a full bank licence granted on 30 December 2020 under Standard Chartered's Significantly Rooted Foreign Bank privileges, rather than a digital full bank licence; Standard Chartered owns 60% and an NTUC Enterprise/FairPrice Group vehicle 40%.
- From September 2026 part of Standard Chartered's Singapore credit card and personal loan clients move to Trust, under an expanded partnership announced on 31 July 2026.
What is Trust AI (GenAI chatbot) and Trust AI Ask?
Trust AI is a generative AI customer-service chatbot in the Trust app, grounded in current product information and each customer's context, that hands conversations to human agents within the same chat; Trust says it handles nearly 50% of customer chats end to end and cut human-handled chats by half. Trust AI Ask, launched on 8 September 2026, lets customers find and total their card spending by asking in everyday language. Behind them, Trust runs AI agents on Amazon Bedrock AgentCore, mostly using Anthropic's Claude models, to triage IT incidents.[8][11][13][15]
What has Trust Bank done on AI, and when?
- Sep 30, 2026
Claude-based incident agents disclosed[15]
Computer Weekly reports Trust's AI agents on Amazon Bedrock AgentCore, using Anthropic's Claude models, cut incident triage to about two minutes; change-review agents are being deployed next.
- Sep 8, 2026
- Jul 31, 2026
Standard Chartered expands investment in Trust[12]
Standard Chartered will invest further in Trust and transfer part of its Singapore credit card and personal loan clients to the bank from September 2026.
- May 19, 2026
GenAI chatbot results published[11]
Trust says its GenAI chatbot cut customer chats handled by human agents by 50% and complaints fell about 40% within months; one former customer service agent now works full time as an AI analyst.
- Apr 30, 2026
- Jul 22, 2025
FY2024 loss narrows to S$93.4 million[7]
The Straits Times reports Trust's 2024 net loss narrowed to S$93.4 million from S$128.4 million.
- Mar 26, 2025
Trust Lock anti-scam tool[6]
Trust launches Trust Lock, letting customers lock Savings Pots and unlock them in the app by tapping their physical card on an NFC phone plus a PIN, with a 12-hour delay; it is a control, not an AI feature.
- Feb 21, 2025
One millionth customer[5]
Trust welcomes its one-millionth customer and says it has become Singapore's fourth-largest retail bank by customer numbers, banking one in five adults.
- May 29, 2023
Deposits pass S$1 billion[4]
Trust says deposits exceed S$1 billion and sets goals to become Singapore's fourth-largest retail bank by customers by end-2024 and to break even in 2025.
- Nov 29, 2022
Thought Machine core banking disclosed[3]
Thought Machine says Trust runs on its cloud-native Vault Core platform and reached 100,000 customers in 10 days and more than 300,000 within two months of launch.
- Sep 1, 2022
Trust launches[2]
Trust launches with a numberless credit and debit card, a savings account and family accident insurance, integrated with FairPrice Group's Link Rewards; shareholders have invested S$400 million.
- Sep 6, 2021
Standard Chartered and NTUC sign joint venture[1]
Standard Chartered Bank (Singapore) and BetaPlus, controlled by NTUC Enterprise, agree to hold 60% and 40% of the digital bank and contribute S$240 million; it becomes Standard Chartered's second separately licensed digital bank in Asia after Mox in Hong Kong.
- Dec 30, 2020
Full bank licence granted[1]
MAS grants SC Bank Solutions (Singapore) Limited, later renamed Trust Bank Singapore Limited, a full banking licence, using Standard Chartered's enhanced Significantly Rooted Foreign Bank privileges.
Where does Trust Bank use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| Trust AI GenAI chatbot[8][11] | In-app generative AI chatbot grounded in product information and customer context, with handover to human agents inside the same conversation; handles nearly 50% of chats end to end. | In production | Customer-facing chatbots |
| Trust AI Ask[13][14] | Natural-language search over card transactions that recognises merchants, categories, dates and amounts and calculates totals; Trust says it filters sensitive data and uses transaction data only to answer the question. | In production | Customer-facing chatbots |
| Real-time sentiment analysis[8][11] | AI sentiment analysis of every customer interaction in real time, which Trust says moved it from complaint handling to complaint prevention. | In production | Customer-facing chatbots |
| Incident triage agents[15] | Agents on Amazon Bedrock AgentCore follow runbooks, pull logs and post a likely root cause with reasoning to the ticket before engineers join; about 65% of analyses are actionable and humans decide the fix. | In production | Generative & agentic AI |
| Change review agents[15] | Agents that check every service a new feature touches before production, trace changes back to requirements and check SRE basics such as indexes and rate limits. | Rolling out | Generative & agentic AI |
| AI backbone for the software lifecycle[15] | Business teams use AI to write machine-readable requirements, metrics, runbooks and remediation steps that agents later use during incidents; Trust says AI has sped up software development. | In production | AI-generated code & coding agents |
| AI-generated AML reports[15] | An anti-money laundering system launched in 2025 in which conventional engineering prepares the data and AI generates a structured report at the end. | In production | AML / KYC |
| Anthropic Claude via Amazon Bedrock[15] | Claude Opus 5 for reasoning and Claude Haiku for lower-order tasks in incident response; most models come from Anthropic, with OpenAI and open-source models under test. Each agent goes to an architecture board with model costs and benefits. | In production | Third-party & vendor AI |
| AWS data lake and Amazon SageMaker[16][3] | Trust is built on AWS, runs Thought Machine core banking on Amazon EKS, tokenises sensitive data in an S3 data lake and is deploying machine learning on SageMaker for personalised recommendations. | Announced | Third-party & vendor AI |
| Trust Lock (non-AI anti-scam control)[6] | Customers lock Savings Pots so funds cannot be moved even if a scammer controls their phone; unlocking needs the physical card and a PIN, with a 12-hour delay. | In production | Fraud detection |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
What has Trust Bank disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Customers | More than 1 million[12] | Jul 31, 2026 |
| FY2025 net loss | S$54 million (FY2024: S$93 million)[9][10] | Dec 31, 2025 |
| FY2025 revenue | S$135 million (FY2024: S$97 million)[9] | Dec 31, 2025 |
| Customer deposits | S$4 billion (FY2024: S$3.8 billion)[9] | Dec 31, 2025 |
| Loans and advances | S$1.2 billion (FY2024: S$800 million)[9] | Dec 31, 2025 |
| Customer chats handled end to end by the GenAI chatbot | Nearly 50%[8] | Apr 30, 2026 |
| Fall in complaints after chatbot rollout | About 40%[11] | May 19, 2026 |
| Incident triage time with AI agents | About 2 minutes (from 15–20 minutes)[15] | Sep 30, 2026 |
| 2025 revenue growth vs cost change | Revenue +39%, costs −7%[8] | Dec 31, 2025 |
“This is not an innovation exercise for us. This is a survival exercise for us.”[15]
Srinivas Patil · Chief Technology Officer, Trust Bank · Sep 30, 2026
“We saw a clear opportunity to redesign customer support with generative AI, not simply to automate conversations, but to deliver a genuinely better experience.”[11]
Angela Yeo · Head of Customer Service, Trust Bank · May 19, 2026
“With Trust AI Ask, customers can simply ask a question about the way they naturally think and speak and get the answer they need in seconds.”[13]
Naveen Sethia · Chief Customer Officer, Trust Bank · Sep 8, 2026
“Profitability was never the end goal in itself.”[8]
Dwaipayan Sadhu · CEO, Trust Bank · Apr 30, 2026
Who runs AI at Trust Bank?
Which regulators govern Trust Bank's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| MAS | Trust Bank Singapore Limited holds a full bank licence from MAS, granted in December 2020 under Standard Chartered's enhanced SRFB privileges. Its GenAI chatbot, Trust AI Ask and Claude-based operations agents fall under MAS's Guidelines on AI Risk Management, finalised on 7 October 2026, and its circulars on generative AI and deepfake cyber risks. | MAS Guidelines on Artificial Intelligence Risk Management MAS FEAT Principles MAS Circular MAS/TCRS/2024/05 |
| UK (BoE / PRA / FCA) | Standard Chartered PLC, which owns 60% of Trust and consolidates it, is a UK-incorporated, London-listed banking group supervised by the PRA; Trust is covered by the group's consolidated supervision, not by a UK licence of its own. | |
| PDPC Singapore | Trust AI Ask and the GenAI chatbot process customers' transaction data and context, which brings them under Singapore's Personal Data Protection Act; Trust says it filters sensitive data and uses transaction data only to answer the customer's question. |
Linked authorities have a page in the Regulation tracker; the others are named for completeness and not yet tracked.
What does the public record suggest about Trust Bank's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
AI is tied to the profitability story
Trust credits automation and early AI adoption for costs falling 7% while revenue grew 39% in 2025, and turned profitable in March 2026. The company has not broken out how much of the cost cut came from AI.
02
Recommend, don't act
The incident agents post a root cause and reasoning but do not take action, and runbooks are tested against a golden dataset. That conservative design matches the human-oversight expectations in MAS's 2026 AI risk guidelines.
03
Anthropic is the main model supplier
Most of Trust's agent models come from Anthropic via Amazon Bedrock, on infrastructure built entirely on AWS. Bedrock's single API makes switching easier, but the current stack concentrates model and cloud dependence.
04
Chatbot savings came with role changes, not stated job cuts
Trust says human-handled chats halved and one customer service agent became a full-time AI analyst; it has not disclosed customer service headcount before or after the rollout.
05
Licensing differs from its digital-bank rivals
Unlike GXS Bank and MariBank, Trust holds a full bank licence obtained through Standard Chartered's SRFB status, and sits inside a UK-supervised group. That gives it a wider product licence and a parent with an existing model-risk framework.
How does Trust Bank use AI?
Trust uses generative AI mainly in customer service and IT operations. Its Trust AI chatbot handles nearly 50% of customer chats end to end and hands harder ones to human agents in the same conversation; Trust AI Ask, launched in September 2026, answers questions about card spending. Behind the scenes, AI agents on Amazon Bedrock triage IT incidents in about two minutes. Trust credits AI partly for reaching profitability in March 2026.
Does Trust Bank use Anthropic Claude or OpenAI?
Mostly Anthropic. CTO Srinivas Patil told Computer Weekly in September 2026 that Trust's incident agents use Claude Opus 5 for reasoning and Claude Haiku for simpler tasks via Amazon Bedrock, that most of its models come from Anthropic, and that it is experimenting with OpenAI and open-source models. The agents cut incident triage from 15–20 minutes to about two.
What is Trust AI Ask?
Trust AI Ask is a feature in the Trust app, launched on 8 September 2026, that lets customers find and understand card transactions by asking in everyday language, such as how much they spent on groceries last month. It recognises merchants, categories, dates and amounts and calculates totals. Trust calls it a market first and plans to extend it to accounts, transfers and rewards.
Is Trust Bank profitable?
Trust reached profitability in March 2026, just over three years after its September 2022 launch, though it still reported a loss for the first quarter of 2026. It lost S$54 million in 2025, down from S$93 million in 2024, on revenue of S$135 million, and held S$4 billion of customer deposits at the end of 2025.
Is Trust Bank a digital bank?
Trust operates as an app-only digital bank, but its licence differs from GXS Bank's and MariBank's. MAS granted it a full bank licence on 30 December 2020 under Standard Chartered's enhanced Significantly Rooted Foreign Bank privileges, rather than a digital full bank licence. It has more than 1 million customers and is supervised by the Monetary Authority of Singapore.
Who owns Trust Bank?
Standard Chartered Bank (Singapore) owns 60% of Trust Bank Singapore Limited and BetaPlus, a holding company owned by NTUC Enterprise and FairPrice Group, owns 40%. Shareholders had invested S$400 million by launch. In July 2026 Standard Chartered agreed to invest further and move part of its credit card and personal loan clients to Trust, which has over 1 million customers.
Who regulates Trust Bank?
The Monetary Authority of Singapore licenses and supervises Trust Bank. Its AI systems, including the GenAI chatbot and Claude-based operations agents, fall under MAS's Guidelines on AI Risk Management finalised on 7 October 2026. As a consolidated subsidiary of Standard Chartered PLC, it is also within the UK PRA's group-level supervision of Standard Chartered.
- Joint Venture in respect of SC Bank Solutions (Singapore) Limited (now known as Trust Bank Singapore Limited)
- Trust Bank launches as the digital bank for Singapore
- Trust Bank Selects Thought Machine to Power the First of Singapore's New Wave of Digital Banks
- Our progress and ambition – becoming Singapore's fourth largest retail bank
- Thank you to our 1 million customers!
- Introducing Trust Lock...a market-first to combat scams
- Singapore's digital banks finding their niche in areas like SMEs as they narrow losses in 2024
- Trust makes strong financial progress and reaches profitability
- Trust Bank swings to profit for the first time in March; Q1 still loss-making
- Trust Bank outpaces digital bank rivals due to strong parent ecosystem: Analysts
- How Trust used Gen AI to half its customer query volume
- Standard Chartered and Trust Bank expand partnership to invest for growth in Singapore
- Trust launches market-first AI assistant for card spends
- Singapore banks up AI game in apps with expenses tracking for retail customers
- Trust Bank cuts incident triage time to two minutes with AI agents
- Trust Bank Builds a Scalable and Innovative Digital Bank on AWS (undated page, retrieved 2026-10-24)
Neobank
BNPL & lending
Payments
Card network
Brokerage
Business finance & infrastructure
When Trust Bank moves on AI, you'll read it here first.
when a bank, fintech or regulator moves on AI, it is in the next morning's brief · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning