Talking is the new typing

Over the last few weeks, many of us at Open have been going into booths in the office, not to take a meeting, but to talk to Claude, over voice.

Just last week, OpenAI added a much more sophisticated voice mode to Codex, its agentic coding platform, and it has taken the developer community by storm. We are also seeing the beginnings of new devices without keyboards or screens. Fringe today, but they may be a sign of what is to come. It feels like there is a non-zero chance that talking is about to replace typing.

When it comes to insurance, talking instead of typing isn't actually new

When insurance really matters, people have always picked up the phone. Getting a quote, submitting a claim, a confusing renewal, "am I covered for this?" Nobody had to be taught to call their insurer. It's what humans instinctively do when something is confusing and consequential.

The industry has spent a decade trying to move those conversations into apps and forms. Yet many customers still prefer to call. When a rigid form or a website can't help, people want to ask for help naturally, in their own words. Insurance is complex, and squeezing your situation into a text box is hard. Often you don't know exactly what to ask. You just know something is unclear, and you want to talk it through with someone until it makes sense.

If we look at the conversations happening across the insurance industry today, most of them are already voice-based. It's just that, until now, both voices were human. Increasingly, one of those voices will be AI.

Consumers are starting to trust AI

For that shift to happen, customers have to trust the voice on the other end, and that trust is building. The FCA's Mills Review puts numbers on it: one in five UK adults say they would already let AI act autonomously on their behalf, within limits they set. As that trust grows, people will be more and more comfortable having their questions answered by AI, and will quickly get used to the instant response times that come with it, whether that's at 2pm or 2am.

Our AI Companion has been voice-first from the start

It's part of the thinking behind our AI Companion, which lets customers ask about their cover out loud and get a straight answer. Doing this safely in a regulated market takes more than a good model, which is why Companion runs inside our AI Safety Harness: five independent layers, from real-time guardrails on every message through to automated auditing of every conversation.

This month, Companion went live inside SuperSaveClub Insurance, the digital car insurance broker we launched with MoneySuperMarket for its 2.5 million SuperSaveClub members. Companion sits alongside the journey with context on the member's actual quote options and where they are in the process, so getting an answer is as simple as asking.

If you are exploring voice in your customer experience, we would love to talk. Fittingly, that one still starts with a human.