Many customer-service teams worry that telling customers they’re interacting with AI will drive people away. Twilio’s August 2026 blog post argues that honest, short disclosures usually preserve — and can even protect — engagement, while late, apologetic, or buried disclosures are what actually sabotage conversations.
Key Details
Twilio’s guidance is built on its customer insights and research. Key points from the post include: 76% of consumers expect an AI agent to identify itself at the very start of a conversation; customers want to know whether they are talking to a human or a robot; and the problem for engagement is disclosure that sounds like a disclaimer, arrives late, or reads as an apology rather than clear information.
Twilio recommends that disclosures be short, specific and paired with an exit to a human. Example wording for channels in the blog includes:
Web chat: “Hi, I’m an AI assistant. I can help with orders, returns, and account questions. Type ‘agent’ any time and I’ll bring in a person.”
SMS: “This is an AI assistant from [Brand]. I can check your order status or start a return. Reply HUMAN to reach an agent.”
Voice: “Hi, this is an automated AI agent for [Brand]. I can help with billing and appointments, or say ‘agent’ and I’ll transfer you.” Twilio also advises leaving a pause after the disclosure in voice so callers have time to decide.
The post highlights several disclosure mistakes that harm conversations: burying the disclosure (for example in footers or pre-chat notices), avoiding the word “AI,” apologising for automation, over-explaining how the model works, and disclosing without providing a clear way to escalate. Twilio cites research showing a gap between brands and consumers: 81% of brands say their AI identifies itself immediately, while only 22% of consumers report the same experience.
Who May Be Affected
The guidance is aimed at companies that deploy conversational AI across web chat, SMS and voice channels. It is directly relevant to teams designing customer greetings, IVR prompts, chatbot scripts and escalation rules.
Why It Matters For Voice AI
Twilio warns that voice interactions are particularly sensitive to hedged or late disclosures because callers can’t scan text. The blog recommends a first-sentence disclosure in voice, naming the agent as automated AI, stating what it can handle, and presenting a clear verbal option to reach a human. The post emphasises that failing to offer an exit often produces bypass behaviour: Twilio’s cited research shows 78% of customers have tried to bypass AI agents to get straight to a human, and 63% of consumers want an easy way to escalate to a live person at any time; only 47% of brands provide that route according to the blog.
What To Watch Next
Twilio suggests operational checks to validate whether disclosures are firing and whether wording changes affect containment and escalation rates. The blog points to Twilio Conversation Intelligence as a way to run language operators across transcripts to count disclosure occurrences by channel and conversation turn, and to compare containment and escalation metrics before and after wording changes. The company says most teams find the feared drop-off does not materialise when disclosures are short, clear and paired with an exit.
Sources
Twilio — How to disclose AI to customers without killing the conversation (published 2026-08-31T20:00:00-04:00). The blog links to Twilio research and the Customer Insights Series report “How to Scale Automation Without Losing Customer Confidence.”
Twilio report — How to Scale Automation Without Losing Customer Confidence (as referenced in the blog).
AiDial analysis: The guidance is operational and channel-specific: simple first-turn disclosures that name “AI,” state scope, and provide a clear human escalation path are the core recommendation. Teams measuring disclosure effectiveness should compare transcript-level disclosure rates and escalation/containment metrics before and after changes, as Twilio recommends.
AiDial: For Australian businesses using conversational AI in customer service, Twilio’s recommendations map directly to script design and escalation flows — disclose early, be explicit, limit technical detail, and always provide an easy human option.
This article is general information and not legal advice.
