A start-up backed by Steve Endacott has created a prototype of an artificial intelligence-based travel agent.
Neural Voice has designed AI Emily to help real-life agents engage with repeat customers about their holiday preferences.
During a spoken exchange, AI Emily captures essential information including the preferred destination, departure airport, party type, duration and board basis.
More: Comment: AI’s great leap forward
Endacott, chairman of Neural Voice and former chief executive of On Holiday Group, said the innovation would give agents a “significantly more effective” way of engaging with repeat customers at the ‘dream stage’ of planning a trip.
He said he viewed Neural Voice’s AI creations as “co-pilots” for agents, rather than “replacements”.
“AI tools are great, but they cannot understand customers’ tones or buying indicators, meaning they are poor at closing the sale,” said Endacott.
“We see them as tools to stimulate demand and pre-qualify requirements before the customer is passed to an online booking process or a human agent to close the sale.”
Neural Voice has built a ‘voice’ overlay that can be used in conjunction with large language models – including ChatGPT – and travel business’ knowledge databases.
Customers’ spoken questions are transcribed into text, inserted into the LLM, and then the response is voiced by AI Emily in the same time it takes a human to respond.
The whole conversation is transcribed with a summary so it can be sent back to the customer and automatically added to the travel company’s CRM. This will help the business to create marketing, provide deep links for online booking or set up an appointment with a human agent to close the booking.
Neural Voice chief executive Jeremy Smith said: “Emily is just the first prototype to demonstrate a concept.
“We are already conversing with some major travel brands to connect to their CRMs so we can inject previous booking information into the conversation to contextualise and personalise it.”
Neural Voice believes its AI agents could also be used by airlines and online travel agencies to resolve most customer service queries.
Smith said: “On a per-minute basis, we are 60% cheaper than human agents on average, so we offer major cost savings”.