---
title: "AI-powered agents for contact center in Australia: containment quality, not deflection"
description: "The Commonwealth Bank episode in 2025 demonstrated that a deflection metric is not a resolution metric: the bank reversed 45 redundancies after call volumes rose despite a voice bot credited with diverting 2,000 calls weekly. By May 2026, on a rebuilt platform, the same automation program achieved 84.6% end-to-end resolution for self-service messaging interactions. A peer-reviewed 2024 study found Whisper’s match error rate significantly higher for Australian English than American English, an effect of roughly 1.3 percentage points, with non-native accents worst affected. Australian regulation as of September 2026 imposes no affirmative AI disclosure duty, but misleading conduct provisions apply and an automated decision-making transparency obligation commences December 2026. Twilio’s platform offers an Australian region with emergency calling, model-agnostic agent execution, and a memory layer that preserves context during handoffs, addressing the containment quality gap."
---

# AI-powered agents for contact center in Australia: containment quality, not deflection

[Twilio](https://yomu.fyi/company/twilio) · Aliakbar Lokhandwala · Sep 28, 2026

**Type:** Explainer

## Summary

The Commonwealth Bank episode in 2025 demonstrated that a deflection metric is not a resolution metric: the bank reversed 45 redundancies after call volumes rose despite a voice bot credited with diverting 2,000 calls weekly. By May 2026, on a rebuilt platform, the same automation program achieved 84.6% end-to-end resolution for self-service messaging interactions. A peer-reviewed 2024 study found Whisper’s match error rate significantly higher for Australian English than American English, an effect of roughly 1.3 percentage points, with non-native accents worst affected. Australian regulation as of September 2026 imposes no affirmative AI disclosure duty, but misleading conduct provisions apply and an automated decision-making transparency obligation commences December 2026. Twilio’s platform offers an Australian region with emergency calling, model-agnostic agent execution, and a memory layer that preserves context during handoffs, addressing the containment quality gap.

## Context

The Australian contact center market is early in AI agent adoption, with 13% of service organizations using AI agents and 17% expecting deployment within six months. The Commonwealth Bank case in 2025-2026 provides a public experiment showing that deflection metrics can mask unresolved calls, while a rebuilt platform achieved 84.6% resolution. Accent recognition accuracy, regulatory obligations, and platform architecture are key considerations for deployment.

## Approach / What changed

Twilio provides an Australian region with emergency calling, model-agnostic agent execution (Twilio Agent Connect) supporting OpenAI, Anthropic, Azure, and AWS Bedrock, Conversation Memory for context during handoffs, Conversation Relay with sub-0.5-second latency and Deepgram Flux turn detection, and Enterprise Knowledge for grounding answers in policies.

## Takeaways

- Deflection rate is not resolution rate; the Commonwealth Bank case showed that a bot credited with diverting 2,000 calls weekly coincided with rising call volumes, and only after rebuilding did 84.6% end-to-end resolution occur.
- Speech recognition accuracy for Australian English is measurably lower than for American English, with a 1.3 percentage point higher match error rate, and non-native accents are worst affected; designing for failure matters more than avoiding the technology.
- Australian regulation as of September 2026 does not require AI disclosure, but misleading conduct provisions apply and an automated decision-making transparency obligation starts December 2026; supervision costs are often underestimated, with 63% of AI-using teams spending more time supervising than handling routine requests.

## FAQ

### What is the difference between deflection rate and containment quality?

Deflection counts conversations that did not reach an agent, while containment quality counts conversations actually resolved. The Commonwealth Bank case showed that a bot credited with diverting 2,000 calls weekly coincided with rising call volumes, because a contained conversation that produces a callback is not a resolved one.

### Does an AI agent have to tell an Australian customer it is AI?

No affirmative disclosure obligation exists as of September 2026. The government's final report on AI and the Australian Consumer Law, published 31 October 2025, found the existing framework fit for purpose. Misleading conduct provisions still apply if an agent actively represents itself as human.

### How accurate is speech recognition on Australian accents?

A peer-reviewed 2024 study in JASA Express Letters found Whisper's match error rate significantly higher for Australian English than American English at p = 0.016, an effect of roughly 1.3 percentage points, with Vietnamese and Thai accented speakers worst affected among non-native groups.

**Tags:** [AI Agents](https://yomu.fyi/topic/ai-agents), [Australia](https://yomu.fyi/topic/australia), [Contact Center](https://yomu.fyi/topic/contact-center), [Speech Recognition](https://yomu.fyi/topic/speech-recognition), [Twilio](https://yomu.fyi/topic/twilio)

- Source: [Twilio](https://www.twilio.com/en-us/blog/insights/AI-Powered-Agents-for-Contact-Centers-Australia)
- Source URL: https://www.twilio.com/en-us/blog/insights/AI-Powered-Agents-for-Contact-Centers-Australia
- Ingested by Yomu: 2026-10-10T06:01:13.646Z

[Read original post](https://www.twilio.com/en-us/blog/insights/AI-Powered-Agents-for-Contact-Centers-Australia)