Presented by Coveo

---

Generative AI has unlocked new possibilities for customer and employee experiences — and as a result, customer and employee expectations have grown. Customers want solutions that enable flawless self-service and offer immediate and accurate advice, while employees want technology that lets them deliver better, more efficient experiences and support them in gaining stronger proficiency behind the scenes.

It can be a trickier proposition than it seems, says Bonnie Chase, senior director, service product marketing at Coveo.

“There’s a lot of excitement about the technology of generative AI, but not much talk about the experience,” Chase says. “It’s key to remember why chatbots weren’t successful in the end — we weren’t able to achieve a consistent and effortless experience that allowed customers to find the information that they need.”

But Coveo has nailed the experience angle, and boasts a 20% boost in a company’s customer satisfaction score (CSAT), and a drop of 30% in cost-to-serve. Here’s a look at what sets an enterprise generative search solution apart.

The promise of a true enterprise generative answering solution

A generative answering solution in the enterprise requires a foundation of secure connectivity, indexing, search and AI, making it possible to personalize anonymous and authenticated sessions based on user context, behavior and intent detection mechanisms along entire digital journeys — all while managing content access via permissions, Chase says.

“A mature solution should embed generative answering across the customer experience, allowing enterprises to generate answers across multiple touchpoints, from any source of structured or unstructured content, all with one AI platform,” she explains.

In other words, generative AI solutions shouldn’t be bolted in as a silo, but threaded into every customer touchpoint. Not only should the customer have a consistent experience, but every employee, from the contact center to the customer success manager, should have access to the same content, same answers and same experiences, as permissions allow.

“That’s what we’re trying to do here: make sure that no matter how customers interact with you, whether they’re searching your website or community or wherever that may be, they’re getting the same goodness that your internal teams get as well, while respecting privacy and security permissions,” she adds.

Shopping for an enterprise gen AI solutions

“The foundation of any digital experience is information,” Chase says. Whether you’re looking to buy, to solve, to explore, it’s all based on content. No matter what you do, no matter how fancy the experience is, if people aren’t able to find the information they need, they’re not going to solve, buy or learn. That’s why the generative answering part of gen AI is crucial for any solution.”

The importance of a unified index

Generative answering is a sophisticated way to interact with search, which relies on how the content is being indexed and how those queries are understood by the search system. Because generative answering specifically relies on existing content to generate those answers, the more access to content that the index has, the better and more accurate the answers can be, which is key for enterprise solutions.

“Our belief is that a unified approach ensures that the generative answering model has the most comprehensive and consistent answers to make it easier for people to get straight to the heart of what they’re looking for,” she says. “That means self-serve experiences like in-product help, documentation and community, in a connected generative experience across multiple touchpoints and secure sources of content.”

A unified index can layer machine learning capabilities on top of the content, which helps automatically optimize each user’s experience based on their interactions with the content. Because it all happens on one index, that optimization is holistic across all content in a system.

Security, permissions and relevance

Enterprises also need to ensure that permissions and security are built into a solution. It should provide answers only based on the content the user has access to. That requires secure content retrieval, basing answers only on what that user has permissions to access, but providing the most relevant answer.

A search platform like Coveo’s Relevance Generative Answering works behind the scenes to create relevance before an answer is generated. The solution integrates LLM technology with the platform to feed generative AI with a common, secure unified index and real-time content and grounded embeddings, helps to drive relevance at scale and consistent factuality, with secure and traceable sources of truth across all channels.

When it’s queried, it identifies the most relevant content to use as the raw data for the answer by searching the index for documents relevant to the question, and then determining the relationships between each document, as well as the relationship between the content in the document and where the document resides in the index.

From there, it creates the prompt sent to the large language model (LLM) which generates the answer from the content and prompt that the solution surfaced, and then that prompt is grounded, so it’s secure and the solution generates an answer within the search page, along with the search results. Not only does the user get a generated answer, but they’re also provided a list of search results and citations, so that the user can self-validate as well.

“This process reduces the chance of hallucinations, making sure that it’s accurate and secure as we provide those answers, and keeps the human in the loop to validate answers,” she explains.

The importance of personalization

Personalization also ensures that users get content that is relevant to their own experiences and needs. A new user looking at onboarding and how-to content will get further recommendations for information that will help them along their journey. A super-user, asking technical questions, will get answers that match the level of their sophistication and their needs.

“Personalization is about making the experience effortless for the customer and for the employees when we’re talking total experience,” she says. It’s ensuring that not only do they have the answer that’s relevant to them, but it meets what they need to move forward with the task at hand.”

Tech stack integration

A generative answering solution that doesn’t have access to all your content isn’t offering you accurate answers. Coveo is currently the most mature solution that unifies content across internal and external sources in a single index. This means no matter where a user submits a query across a company’s tools, the answer will be derived from every source of information it generates, Chase explains, including content such as help videos on YouTube, or social media conversations.

“Companies don’t just have one place where they’re creating knowledge,” she says. “It’s important for those to be indexed as well, because the best answer isn’t always in a document. And the unified index ensures consistency no matter where you are, no matter what tech stack you’re using or how many tools.”

Relevance Generative Answering in action

Coveo’s Relevance Generative Answering is live and producing results right now, Chase says. The platform is already differentiated by its sophisticated applied AI, designed to deliver highly relevant, bespoke digital experiences that drive superior business outcomes. Its scalability, rapid time to value, enterprise-grade security and compliance, and native integrations with other third-party technology applications are also unique.

And still, the company is continuously testing, iterating and evolving, working with 20 design partners and beta testers, including Informatica, Synopsis, VMare, Xero and Zoom, as well as 25 additional customers in its advisory group. At Salesforce’s recent Dreamforce conference, Xero shared that its partnership with Coveo is now enabling them to personalize the experience for millions of customers across different touchpoints and platforms, and helping them transform the customer service experience from Reactive to Proactive.

Coveo also serves as “customer zero,” using the technology internally as a proof point for the technology’s effectiveness in four places: documentation, community, in-product experience and the internal help desk. The aim is to build a case for the enterprises they talk to that don’t quite trust the technology yet. These leaders know they need to adopt it, but they’re not sure where to start, Chase says.

“By using our technology ourselves, with employees and customers, we are showcasing that it’s trustworthy,” she explains. “And we’ve helped hundreds of the world’s leading brands create tangible financial value over the last decade.”

Dig deeper: Download the ebook “GenAI Headaches: The Cure for CIOs” to learn how you can stay ahead in the quickly-developing AI landscape. 

---

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact [email protected].