Good Knowledge Is the Basis for Good AI: eGain Says at Solve Event

CHICAGO -- Poor quality at the knowledge layer is why there's an artificial intelligence value gap within most organizations, eGain CEO Ashu Roy told the audience during his keynote address at the eGain Solve 26 conference today.

A few figures demonstrate the extent of the value gap: AI investment and executive expectations are rising faster than measurable results, 91 percent of customer service leaders face pressure to implement AI, service AI spending increased 38 percent, while service budgets rose only 2 percent.

However, there are some areas where the investment in AI is paying off: Companies are achieving positive ROI in about one quarter of 400 service AI use cases, compared to results for the remaining three quarters, which were negative, break-even, or unknown, including 41 percent classified as unknown.

Customer-facing AI is the highest-risk use case, according to Roy, who said users cannot control inputs or judge output quality. Trusted inputs, guardrails, and continuous evaluation are required to reduce harmful false positives, he added.

Knowledge maturity requires high quality and relevance but cannot be maintained manually at scale, Roy added. AI can automate sourcing, creation, management, structuring, and verification of human-consumable knowledge.

Knowledge has been the basis of eGain's business ever since the company was founded. "AI helps us break through that long journey, and that's what we are excited about," Roy said. "Interestingly, we announced our first product, Instant Answers, four years ago, right around this time, just before the introduction of ChatGPT."

AI KnowledgeOps Introduction

eGain used the confernence to unveil AI KnowledgeOps, which eGain describes as an industry framework for keeping enterprise knowledge accurate and ready for AI.

Trusted AI starts with knowledge, according to Roy. Company research shows that around 80 percent of a knowledge base's value tends to come from about 20 percent of its content. But that content is often scattered across several systems.

AI KnowledgeOps starts with the company conversations happening across every channel and the content spread across its systems and uses both to uncover the questions that customers are asking and which topics matter most to the business. Those questions are checked against existing content to find what's missing or incomplete, so teams can close the highest-impact gaps first. Knowledge is tagged and mapped to the appropriate products, policies, and rules so AI can find the right answer for the right person.

Four new eGain capabilities help teams drive a more efficient and accurate AI Knowledge Ops process:

• WYCA (What Your Customers Ask) is a free tool that reads customer conversations across channels and groups together the many ways people ask the same question. It then shows which questions matter most to the business because they matter most to customers.
• eGain Knowledge Graphs provide AI agents with a map and a shared vocabulary of how businesses operate.
• eGain Genie is a conversational AI agent built into all eGain products.
• eGain kForge is a services accelerator that connects every content source and eliminates the painful migration typically associated with getting content ready for AI.

"The first thing every executive wants to know before putting AI in front of customers is whether it will get the answer right and take the right action, and both come down to the knowledge behind it," Roy said. "AI KnowledgeOps gives companies a repeatable process to make sure it does, even as their business keeps changing. We've spent more than 25 years helping enterprises manage knowledge, and the organizations seeing real results from AI are the ones that treat knowledge as an ongoing practice. The capabilities we're introducing today make that practice faster and easier to maintain, so teams can bring their content together, see what customers actually need, and keep every answer and action current."

eGain also introduced the eGain Trusted Knowledge Network with prebuilt connections to more than 100 national and regional regulators and other authoritative industry sources across financial services, insurance, and healthcare. Workflows can draw on the Trusted Knowledge Network and use AI Knowledge Connectors to execute actions across multiple systems, so an AI agent can handle requests from start to finish.

eGain industry AI Knowledge Connectors now include new integrations to allow banks, insurers, and healthcare organizations to exchange governed knowledge between eGain and other systems. eGain now integrates with Jack Henry core platforms for banks and credit unions, adding to its existing industry connectors.

Formula for AI Success

AI success depends less on model adoption than on the maturity of the knowledge foundation beneath it, said Amy Machado, IDC research director.

IDC research found that organizations with the strongest knowledge foundations enjoyed 85 percent more business impact from AI knowledge-management investments than organizations at the low end of the maturity scale.

Ninety-five percent of organizations are evaluating, piloting, or deploying AI in knowledge management, making adoption itself a weak differentiator, according to Machado, who shared the following three pillars for AI success:

• Technology: Leaders use unified knowledge platforms as a single source of truth for employees, AI, and agentic systems. Fifty percent of leaders use a unified platform, compared with 5 percent of laggards.
• Process: Leaders automate content capture, tagging, routing, reconciliation, personalization, and publishing. Ninety-four percent of leaders automate more than half of their knowledge-management processes, compared with 36 percent of laggards.
• People and structure: Leaders centralize governance, assign accountable ownership, and develop knowledge-engineering capabilities such as ontologies and shared enterprise vocabularies. Thirty-one percent of leaders have a centralized knowledge-management structure, compared with 6 percent of laggards.