NLX Debuts Generative AI Conversation Builder for Customer Service
Customer service AI developer NLX has brought generative AI to its conversational designer for virtual assistants. Businesses can use the new feature to quickly build and deploy voice and text-based agents on the Conversations by NLX platform just by describing the purpose of the conversation for the bot.
NLX Generative Conversation
Enterprise services are keen to use generative AI, and the conversational capabilities of ChatGPT and other LLM-powered services are undergoing a lot of testing and refinement for customer service and other business needs. But, concerns about hallucinations and incorrect or inappropriate responses by a generative AI chatbot inspired NLX to come up with a way of leveraging the technology to help create more reliable conversational AI experiences, as seen in the video above.
NLX envisions the generative AI feature as a way to speed up the planning and implementation of conversational flows for companies using automated chatbots or voice agents. Instead of writing an entire script, the user can describe the point of the conversation in broad terms, and the AI responds like an actor improvising a scene to meet that goal. The interactive creation continues throughout the process, with the AI accommodating more scenarios, specific requirements, and circumstantial changes. An initial question-and-response can grow into a complex web of responses able to handle whatever a future customer might say.
“Business processes are structured in nature and a pure generative text model is unfit to handle them due to the statistical nature of machine learning,” NLX CEO Andrei Papancea told Voicebot in an email. “For that reason, conversation building, especially for structured flows is still a manual process — unless you’re using NLX’s new AI coach for accelerating conversation building.”
This isn’t NLX’s first foray incorporating generative AI. The company started embedding OpenAI’s GPT-3 LLM into the Conversations by NLX platform alongside Dialogflow and Amazon Lex back in February. The GPT-3 package accounts for hallucination concerns with extra restriction options, but that reduces its flexibility in other ways. NLX counts airlines, banks, hotels, and other businesses among its clients, all of which might be interested in streamlining the design of their automated conversation services.
Contact centers are experimenting at a furious pace with generative AI and the trend seems to be accelerating. For instance, contact center AI platform Observe.AI recently debuted its own 30-billion-parameter LLM and related generative AI tools. And customer service automation startups like Ada, NLX, Hyro, and Conversica have all embedded generative AI in some form into their platforms. NLX’s idea to widen the scope of generative AI’s place to the behind-the-scenes composition of conversations might entice companies wary of letting the AI loose in their systems, and points toward a possible future with broader integration of generative AI by enterprise service providers.