Voice AI is bringing the phone call back to business

AI voice has moved ahead by leaps and bounds. Voices are no longer robotic, and the rise of sophisticated language models and agentic systems enables voice AI to answer questions, pre-empt customer queries, and deliver fast and accurate information quicker than many human agents.

Over recent years, many organisations have tried to shift customers away from using phone calls for service as it can be time-consuming and costly. Anyone who has looked for a company phone number has seen the attempts to shift callers away from physical agents to chatbots, FAQs and other forms of online help. But that is changing.

Robert Woolfrey, the Vice President of Sales for APJ at Twilio explained.

"What I see is customers no longer wanting to bury a phone number. In the past, you would take a phone number and make it as inaccessible on your website as possible to try to deflect from people calling in and drive them more towards email or chat. Today, we're seeing a surge in voice. Customers are adding phone lines."

Woolfrey said the surge in voice is broken down into two main categories: sales and customer care. But the use-cases are not limited to externally facing functions. Internal workflows are also seeing the benefits. Processes inside a warehouse are used to trigger calls to customers. When an order is picked, packed and dispatched, the customer receives a voice call telling them the status of their order.

Twilio is using these tools internally. When a potential new customer calls, a voice-powered AI agent answers and guides them through the call.

The use of voice has many other applications. As Chris Connolly, Twilio's Director of Solution Engineering for APJ, explained.

"We did a lot of research with a customer in Australia that has brokerage in its portfolio. Their problem is that they work on a million-dollar loan to a customer and then go see someone else. But they need time to write notes and get them into the CRM. So we created an experience as part of a hackathon to build that. They now have a hands-free AI experience."

Connolly demonstrated a voice assistant that that could receive a call from a customer and help them troubleshoot and resolve a settlement issue with an EFTPOS machine. As well as resolving the initial issue, the AI retained history of the call as a memory it could reference in future calls.

And he's seen a similar use-case with field service workers in Australia. They can receive a job, update the details, and have it all sent into the ITSM system.

"This is a perfect use case for voice," added Connolly.

Twilio's voice platform offers significant flexibility. It works with a variety of partners including DeepGram, Google, Eleven Labs and Amazon. But customers can integrate their LLM of choice. And with different applications offering support for a variety of languages and regional accents, something Woolfrey said is important to remember when creating a voice solution, it's possible to use the platform to solve very specific business problems.

One of the biggest challenges organisations face on their AI journey is scaling from an initial project or proof of concept to production. Connolly said this is not an issue with Twilio as the company was founded almost two decades ago with scalability as part of its DNA.

"The elasticity of scaling from zero to millions is the problem we solve," said Connolly. "You can't vibe code scale."

This is a critical consideration as contact centres are transforming. Contact Centre as a Service (CCaaS) has leaned heavily on carriers for many years. But with the ability to create a CCaaS user interface that fits a business more easily and better than ever before, scaling becomes a more immediate challenge.

"This is where we think Twilio comes into its own," said Connolly. "We've spent years building APIs and interfaces at scale so that businesses from a sole operator all the way to the largest banks in the world can use the exact same platform and get going immediately."

Already, Twilio is seeing new companies scale quickly by leveraging these new tools and competing with much larger competitors. Organisations using new platforms are unencumbered by older legacy systems and can offer a full range of products and services with greater agility.

"Businesses that are adopting AI aren't communicating less with their customers, they're communicating more," said Woolfrey. They might not be talking to the contact centre, but that doesn't mean they're not talking to their customers. In every single customer service use-case, AI doesn't mean you communicate less, you communicate more."

As voice AI becomes more capable, natural and scalable, the phone is shifting from a costly service channel into a flexible tool for customer engagement and internal operations. The opportunity lies not simply in automating conversations, but in combining voice, context and business systems to deliver faster and more personalised experiences at scale.