Yext’s Foray into Conversational AI for Enterprises
Move over, Google. Yext has launched an AI-powered chatbot in closed beta. Known as Yext Chat, it lets enterprises integrate a white-labeled chat function on their web properties. This can be for a variety of customer service use cases.
If this sounds familiar, it borrows from the Yext Search playbook, which offers a customized search engine to any website. Like Yext Search utilizes a given enterprise’s knowledge base as a search index, Chat will tap into a similar set of resources to drive its AI engine, training set, and language models.
In both cases, Yext taps into the same technology: the Yext Knowledge Graph. This is sort of a customized and site-specific version of Google’s Knowledge Graph. Yext Chat will utilize this, along with multiple large language models including GPT-3, and manual controls for enterprise customization.
“The world is waking up to the power of large language models that can generate truly coherent conversations,” said Yext Chief Data Officer Christian Ward “We’ve built Yext Chat on top of a knowledge graph to bring these capabilities to the enterprise in a way that’s compliant and authoritative. In order to effectively use conversational AI within a local business context, every organization — no matter how big or small — will need a centralized, curated set of knowledge to ensure that generated answers are accurate and grounded in real-world information. A knowledge graph enables this approach, all while future-proofing an organization’s content for emerging technologies that will leverage different models for different objectives.”
B2B2C Play
Besides enterprise customization, Yext Chat is advantaged by its agnosticism. It’s designed to plug into existing enterprise software platforms including ticketing systems, e-Commerce platforms (think: checking order status), or Slack workspaces. This casts a wide net in terms of addressable market.
But the headline for Yext Chat is perhaps its timing. As noted, there’s lots of excitement around generative AI (Dall-E 2) and conversational AI (Chat GPT). But there’s a supply gap where enterprises can’t easily integrate their own personalized conversational AI agents. That’s where Yext Chat comes in.
In that sense, Yext Chat may have beaten Chat GPT to its own market. Open AI’s business model will likely involve a B2B2C play. In other words, consumer-facing enterprises will pay to integrate the technology to serve their customers.
But that’s not to say that Yext is being derivative. As noted, the company has already mastered this same model with Yext Search. Following a similar playbook, Yext Chat will be advantaged – not just in terms of established technical competency, but also in Yext’s existing enterprise sales channels.
Burning Questions
All the above could mean that Yext has a built-in advantage in selling and bundling Yext Chat to existing customers. Those customers currently use a mix of Yext’s core platform for search and business listings optimization, as well as its newer forays like Yext Search. Chat now joins the suite in synergistic ways.
And it does so with impeccable timing. Large language models aren’t only in high demand, but they need lead time to do their thing. The sooner that Yext can get its chatbot deployed at scale, the sooner it can get smarter and carve out greater technological differentiation. That’s the name of the game in AI.
Meanwhile, to address these questions and speculations, Street Fight’s parent company Localogy will have Ward on stage at L23 in April. Perhaps by then, Yext Chat will be smart enough that we can invite it on stage as well to answer our burning questions. Meanwhile, you can go hands-on by signing up for the closed-beta waiting list here.