LivePerson has upgraded its Conversational Cloud platform with generative artificial intelligence and large language model (LLM) capabilities across voice and messaging channels.
LivePerson's AI is trained on a robust data set informed by a billion conversational interactions every month.
"Businesses have always dreamed of automating truly human-like conversations on a massive scale, but the effort and expense put this dream out of reach until the dawn of generative AI and LLMs. But the hard truth is these technologies are not fit for business use right out of the box," said LivePerson's founder and CEO, Rob LoCascio, in a statement. "Our new trustworthy AI capabilities offer guardrails designed to make LLMs safe and effective for even the world's biggest brands, all while bringing digital experiences for their employees and end-consumers to new heights."
The new capabilities in Conversational Cloud allow companies to do the following:
- Restrict LLM responses to the guardrails of a curated collection of knowledge content managed within LivePerson's controlled environment;
- Determine the source of content errors and vet every response an LLM generates; and
- Access safety and usability analytics in a dashboard that includes content moderation.
The new platform also includes the following:
- Conversation Copilot, which can provide recommended answers for human agents based on content in knowledge bases, websites, PDFs, spreadsheets, and more; auto-summarize transferred or escalated conversations; and accelerate bot creation by generating diverse sets of customer training phrases.
- Conversation Autopilot, which directly assists customers with ChatGPT-like conversations, automatically generates no-code virtual assistants from knowledge base content, and provides detailed and natural responses across both voice and digital; and
- Voice AI, which automates expensive phone calls with voicebots; drives customers to the voice or messaging channel best suited to help them based on their intent, sentiment, and needs; and ingests voice calls for analysis within LivePerson's platform.