From AI agents that make calls on your behalf to aged-care outreach calls, Twilio’s programmable voice is emerging as the invisible layer powering voice AI.
When you receive an OTP from your bank, a flight confirmation from an airline, or call a contact centre, there’s a good chance Twilio is quietly powering it, according to Robert Woolfrey, Twilio’s VP of sales in APJ.
“Think of us as a cloud-based infrastructure layer that allows businesses to communicate with their customers across any channel.”
Twllio’s director of presales engineering and solution architect in APJ, Chris Connolly, explained, “What we power is that communication fabric… all of that is part of the Twilio platform. And then where we stop is at that intelligence layer where the large language model (LLM) comes in.”
From CPaaS provider to AI-ready voice fabric
As enterprises race to deploy “voice AI”, Twilio argues that the real foundation is programmable voice – the software layer that can make, receive and route calls, and plug into AI models.
Chris explained, “When we’re talking about speech recognition, understanding what people say, text-to-speech, your synthetic voices… as well as… background noise suppression… all of that is part of the Twilio platform.
Crucially, Twilio is model‑agnostic. “We need to support our customers having choices and bring the LLM that makes the most sense for them in the market they’re trying to work in.”
Self‑service programmable voice
Programmable voice itself predates the current AI boom. “I don’t think it’s because of AI. I think it’s the other way around,” Chris argued.
“A phenomenal amount of self-service customers are able to turn on Twilio, get an account, and in the same minute start to build a voice experience.”
Robert Woolfrey
“If you look back at history, Twilio has been the behemoth in that category for 10 years. It effectively invented CPaaS and programmable voice… What that has allowed is developers to build these new experiences. Now AI workloads are taking advantage of that… layer of programmability to create new experiences.”
That makes Twilio a natural partner for the major AI players like Open AI which Chris noted is so easy for OpenAI to plug into Twilio.
Robert also shared about how easy it is for developers and businesses to start to build on the Twilio platform. “A phenomenal amount of self-service customers are able to turn on Twilio, get an account, and in the same minute start to build a voice experience.”
This has had a democratising effect and as a result Robert shared self-service revenue is growing 30-percent year-on-year now.
Agents that call for you, with consent and guardrails
The flexibility of programmable voice also raises questions on privacy, security and abuse. Twilio’s platform lets developers control call flows, collect keypad/speech inputs, route calls, integrate voice data with other systems and so on.
Twilio’s answer to potential security/privacy concerns and abuse starts with clear labelling.
Chris explained the transparency approach that Twilio takes and AI nutrition facts they share that are “…a way to very clearly understand what is in the solution that you’re offering. Customers stay in control of their data and whether it can be used for training, he added, stressing opt‑in consent.
Twilio also warns against giving agents too much unchecked power – what it calls “excess agency”.
“It is a point where we talk to CISOs and CIOs about how they think through that process. Where is the decision making done for their customers? Anything that touches their life or their profits needs to really be thought through.
“The way that we think about it at Twilio is the companies we work with are ultimately in control of that data, and they control the process, and they’re able to have explainable decisions all the way down through the stack,” he said.
“When you have an agent that is operating on your brand or potentially on behalf of your customer, it needs to have the same controls as a human… You don’t give the keys to the building to a contact centre operator… in the same way, you wouldn’t give your agents keys to everything as well.”
The Japan market
Japan is one of Twilio’s fastest growing voice markets in APAC. Robert shared about the healthy independent software vendor (ISV) market in Japan and how Twilio has watched and helped to scale over the last two years.
Both execs are also on the same page about their impressions of the Japan market – trust-heavy, highly nuanced, with demand for deep localisation.
Chris said, “We operate a local capability in Japan specifically to address some of those concerns, and what that means is that you can use capabilities from our partners or enterprises that can produce their own AI capability, plug into the Twilio infrastructure in Japan, and keep that data local.”
If before, voice-based innovation had to wait for US software companies to come to Japan, Robert sees that there is a lot of opportunity for independent software vendors (ISVs) to build on top of Twilio’s infrastructure layer, something very specific for a Japan market.
Robert concluded, “That local nuance is so important. Probably more so in Japan than you see in some other markets. Other markets are more tolerant of certain misses… but the Japanese are focused on the details.
“It’s going to be really interesting to see how foreign companies perform in that market in this space.”