Commonwealth Bank of Australia
GenAI customer messaging + proactive alerts
In 2024, Commonwealth Bank of Australia's GenAI customer messaging handled more than 50,000 enquiries a day, with a 40% drop in call center wait times and a 50% reduction in scam-related losses.
Key points
- GenAI customer messaging in the app plus proactive suspicious-transaction alerts.
- In-house GenAI layer on the transaction flow, with NameCheck, CallerCheck, and CustomerCheck.
- More than 50,000 enquiries a day, wait times -40%, scam losses -50%.
- Evidence level B, live status confirmed.
Objective
Shift call volume to GenAI messaging in the app to reduce wait times, and use the same AI layer to send proactive alerts that cut fraud and scams before they succeed.
The deployment
Commonwealth Bank put generative AI-powered customer messaging in its mobile app, which handles more than 50,000 retail customer enquiries a day. Over the past financial year, the bank attributes a 40% reduction in call center wait times to this messaging. The same AI component sends proactive alerts on suspicious transactions (around 20,000 a day, with a planned ramp-up to 35,000), which CBA links to a 30% drop in customer-reported frauds and a 50% reduction in scam-related losses through its NameCheck, CallerCheck, and CustomerCheck features.
Results Proof B
Figures published by CommBank in an official communication (T1) aligned with its FY24 results, picked up by specialized press (T2/T4). Operational metrics tied to the results but not audited P&L lines, hence B rather than A.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- llm Couche GenAI interne CommBank generative messaging developed in-house; the underlying models are not publicly disclosed
- outil Detection d'anomalies transactionnelles in-house models that continuously scan the transaction flow and trigger the proactive alerts (about 20,000 a day)
- outil NameCheck / CallerCheck / CustomerCheck in-house anti-scam features to which CBA links the 50% reduction in scam-related losses
- infra App mobile CommBank single channel for the GenAI messaging and the proactive alerts, wired to the core banking system
How it runs, concretely
For ops teams-
1Messaging enquiry AI
The customer writes in the app messaging; the GenAI layer answers common enquiries and relieves the call center.
-
2Transaction monitoring AI
In parallel, the system scans transactions to spot fraud and scam patterns.
-
3Proactive alert AI
On a suspicious transaction, an alert is pushed to the customer (about 20,000 a day), with the NameCheck, CallerCheck, and CustomerCheck controls.
-
4Escalation and management customer service / risk team
Complex or high-risk cases go to an agent; the teams track wait times, fraud rate, and losses to adjust.
Real-time transaction data. It feeds both the messaging responses and the anomaly detection for the alerts. Without a clean transaction flow, the proactive alerts become noise and lose customer trust.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- Real-time transaction flow
- Labeled fraud and scam history
- Knowledge base of frequent customer enquiries
Org prerequisites
- Coordination between customer service and the risk/fraud team
- A proactive messaging policy (frequency, false positives)
- A compliance and AI transparency framework
Possible stack
- Generative LLM for messaging
- Anomaly detection models on transactions
- Mobile app and core banking integration
The plan, step by step
- Step 1Map the most frequent customer messaging enquiries and set the compliance framework (transaction data, AI transparency).Deliverable: Prioritized enquiry corpus and validated legal framework
- Step 2Build the GenAI messaging on the top enquiries, integrated into the app, with an equipped escalation path to an agent.Deliverable: Working internal beta with an escalation flow
- Step 3Roll out progressively to customers and measure the effect on call center wait times.Deliverable: Messaging in production with a deflection dashboard
- Step 4Wire anomaly detection onto the real-time transaction flow and calibrate false-positive thresholds.Deliverable: Alert engine validated in pre-production
- Step 5Launch the proactive alerts with a shared customer service / risk team loop and track reported frauds and losses.Deliverable: Alerts in production with fraud and loss monitoring
First step: Start with messaging on the most frequent enquiries to reduce calls, then wire anomaly detection onto the transaction flow for alerts.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 Customer safety, convenience and recognition boosted by early implementation of Gen AI Primary archive pending
- S2 Commonwealth Bank uses AI to enhance customer service Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.