Content, AI, and knowledge graphs: The key to driving positive customer experiences with TopQuadrant

While enterprise content is the key to developing positive customer experiences with AI, finding and extracting knowledge from this content—wherever it lives, across a diverse set of systems—is a significant obstacle. With content being manually tagged, human error leads to inconsistencies across systems, driving poor discoverability, wasted time, and poor customer experiences.

In KMWorld’s latest webinar, In AI, Content is King. But Finding and Leveraging It Is the Real Challenge, Nimit Mehta, CEO, TopQuadrant, and Laurie Nelsen, semantic solutions engineer, TopQuadrant, explained how through the power of knowledge graphs, a top ecommerce company was able to achieve a robust content management strategy capable of fueling AI technologies.

One of TopQuadrant’s clients—an ecommerce company focused on travel aggregation—established its value by helping customers make travel decisions smarter and faster. Ultimately, the optionality that it toted became its “Achilles’ heel,” according to Mehta, where the processes supporting that optionality became increasingly complex. A simple decision about travel details—what hotel you should stay at, what its ratings are, what amenities it has, etc.—was complicated by a massive web of available information.

“The more data you have, the harder these decisions are becoming,” said Mehta. “All the data required to make a decision is overwhelming…this was primarily challenging for a company that was built around providing as much information as possible to everyone at the click of their fingertips.”

Compounding this was a customer desire for personalization; the ecommerce company found that while more customers look for personalization in their travel aggregator, it is increasingly challenging to provide that level of customization at scale. According to the ecommerce company’s data, they were losing $6.4-7.2 million in revenue per month due to poor personalization, with a 93% drop-off rate due to customer confusion.

Upon completing a root cause analysis, the company found that inconsistent and disconnected content was the technical pain behind their inability to personalize customer experiences at scale. From a growing amount of content to no consistent classification system, tooling limits, and no reconciliation for clashing classification, the company found its weakness—a lack of a KM strategy.

TopQuandrant helped them find a solution: A content classification system that could align the company, acting as a corporate memory that was uniform and expressive, collaborative and governed, and above all, reusable. This single content tag graph would:

• Connect master, reference, and taxonomy data
• Encode business-aligning descriptions to power intelligent tagging
• Offer a GIT-like experience reconciling judgements, updated collaboratively, and propagated everywhere in real time
• Deliver personalized user experiences through aligned content and data
• Enable governance to reuse and power multiple activities around aligned descriptions
• Accelerate the creation of trained AI agents

Ultimately, graphs are the only way to unify content, according to Mehta. After implementing this system, the travel aggregator saw:

• $12.6M-plus revenue in 24Q3
• -24.6% drop-off rate
• 18%-plus KM cost savings through reuse
• 2-plus AI Agents per month, piloted and tested

Nelsen further echoed Mehta’s sentiments, propelled by real-life experiences where a lack of content management strategies prevented Nelsen from effectively finding the resources she needed. From her experiences in library and information science to education and healthcare, Nelsen developed a passion for organization information that brought data consumers closer to the information they needed, when they needed it, to drive positive, impactful outcomes.

Nelsen pointed to knowledge graphs—capable of capturing a web of interdependencies and classifications—as the most adequate framework for content management. The knowledge graph acts as “a central place that dynamically tailors the response. It’s not people going out and looking for things in all these different places, but it can actually respond in real time to what they’re doing,” said Nelsen.

This is only a snippet of the full webinar brought to you by KMWorld. For the full webinar, featuring in-depth explanations, use cases, a Q&A, and more, you can view an archived version of the webinar here.