Nuance has become the latest CX vendor to incorporate large language models into its conversational AI platform: Nuance Mix.
The first use case will see ChatGPT take over a conversation when the consumer asks a question the bot has not been trained to handle.
So the GPT-powered bot has content to fall back on, the creator fuels it with business-specific resources, such as a website link, online manual, or customer-facing knowledge base.
Parsing these and spotlighting relevant insights from the resources, the bot can answer more customer queries.
Finally, the “conversation booster” supports users as they carry out tasks within the bot, like completing a transaction.
The solution detects when the customer changes their intent mid-way through such a process, brings the conversation back on track, and – according to Nuance – improves containment rates.
In doing so, it harnesses APIs from OpenAI. As Tony Lorentzen, SVP & General Manager of Nuance, stated in a recent blog post:
This gives customers access to solutions that use the most advanced open AI through a trusted cloud provider and responsible AI infrastructure, and on a platform based on decades of contact center AI leadership.
A preview of the conversation booster feature is now available to those who sign up for Nuance Mix and join the waitlist.
GPT-Fuelled Innovation and the Conversational AI Space
Conversational AI providers have spent years crafting the NLP models within their platforms to analyze intent, keywords, and phrasing.
Now, large language models – like ChatGPT – are erasing much of that modeling because they do not need training. They extract intent and entities out-the-box.
As such, many vendors may use the technology to level the playing field with market leaders.
Yet, leaders are also harnessing large language models as a vehicle for future innovation – as evident in Nuance’s conversation booster – enabling a greater degree of contact automation.
In most cases, this is with chat as the modality. Yet, from this point forwards, chat, voice, and text are all inextricably linked.
Indeed, there are already engineers attaching synthetic voice to Chat-GPT.
Such moves may spark the fuse for further innovation in the field and shake up the entire conversational AI landscape.
As Bradley Metrock, Founder & CEO of Project Voice 2023, recently told CX Today:
ChatGPT has rallied resources to conversational AI in a remarkable way. Investors and talented engineers are sitting up straighter and paying attention to the space in a way they didn’t before. Its attracting quality to the universe.
Also, it is interesting to consider how these engineers may enhance the low-code/no-code aspect of building conversational AI.
Indeed, in the future, conversational designers may tell the platform what they want, and it then creates a corresponding bot – wiping out all the dragging and dropping in between.
Nuance hopes to benefit from Microsoft’s close relationship with OpenAI and stay on the cusp of such large language model innovation in conversational AI.
But, now ChatGPT has gained enterprise APIs, expect innovations aplenty across the space, which may just turn the conversational AI market on its head.