Pathlight promises ‘human-level’ conversation analysis with new multi-LLM tool

California-based Pathlight, a company providing insights into enterprise and team performance data, today announced the launch of Conversational Intelligence (CI), a tool that leverages several major language models for in-depth analysis of interactions with customers.

Available starting today, Pathlight CI takes a step ahead of traditional machine learning-based solutions and deciphers the nuances of language and context to provide businesses with deep context and trends in conversations with customers. According to the company, it can provide human-level analysis while maintaining the speed and scalability synonymous with AI.

The development highlights another notable use case for large language models in the enterprise setting, as well as popular aspects such as text and image generation.

How exactly does Pathlight CI help?

Interactions with customers are sources of risks and opportunities. Businesses want to leverage all incoming calls, messages, and tickets to define their policies, but are often slowed down by manual quality assurance teams that only review a small portion of conversations, or by traditional intelligence solutions shaped to capture certain keywords and phrases, not the full context of the interaction.

“The most critical business insights and customer issues come from day-to-day interactions, yet most organizations have visibility into less than 5% of conversations. You can’t fix what you can’t see,” Pathlight CEO Alex Kvamme told VentureBeat.

To fill this particular gap, Kvamme and his team came up with CI, which uses transfer learning, transformer architecture, and multiple LLMs to analyze every conversation on every channel in real time and uncover hidden insights.

“Each conversation is summarized, tagged, customer intent and problem resolution are categorized, analyzed against personalized questions, and scored for agent performance,” Kvamme told VentureBeat.

At its core, the tool’s “pluggable” infrastructure allows multiple LLMs to perform different stages of the data processing pipeline, depending on the task at hand. For example, a privately hosted internal model of the company assists in drafting PII before a publicly hosted model performs general summarization and markup and a third specially trained model performs customer and agent analysis. .

Once the analysis is complete, the tool summarizes the results for executives to help them understand exactly what’s going on in their millions of conversations and make decisions accordingly.

No need for fine tuning

While LLMs can be a simple solution for analyzing customer conversations, it’s important to note that existing models aren’t ready for off-the-shelf proprietary business data. They need to be optimized, which is a heavy engineering task. With the dedicated Pathlight CI offering, which primarily requires connecting a conversation source and setting up an AI processing pipeline, this step can be skipped.

“No tool leverages the build AI at the scale and depth of Pathlight CI… Something this simple is easy to build, but due to throughput and token limits, it’s impossible to deploy it at scale. No public LLM will support the analysis of millions of conversations, which is why Pathlight has invested so deeply in its native LLM infrastructure,” Kvamme added.

The company has already tested the offer with several companies, with some already signing up for the paid version of the tool. The CEO did not share the names of those customers.

Previously, Pathlight has worked with companies like Clear, Kia, Stitch Fix and Twilio to power their customer-facing teams.