eGain Links AI Cost Overruns and Chatbot Failures to Poor Information Organization

eGain Corporation has released two whitepapers identifying a common cause for two significant issues plaguing enterprise AI: runaway costs and ineffective AI chatbots. The company asserts that neither problem is inherently about the AI's intelligence but rather the lack of organized, accessible information that AI systems rely upon.

The first paper, "Ending Runaway Token Costs in Enterprise AI," addresses the rapid escalation of AI expenses. eGain explains that a substantial portion of these costs is due to wasted processing. When AI systems lack organized knowledge bases, they must re-process vast amounts of documentation for each query, akin to repeatedly searching an unlabeled warehouse. Additionally, using AI for routine tasks that could be handled by simpler software adds unnecessary expense. eGain proposes organizing company knowledge into a unified, mapped source to streamline AI operations and route rule-based tasks to conventional software.

The second paper, "The Trust Gap in AI Self-Service," examines why AI deployments in customer-facing roles often stall before going live. The core issue, according to eGain, is the risk of AI errors when interacting directly with customers. While 95% accuracy might sound high, the remaining 5% of errors can have significant consequences. The paper also critiques the practice of allowing AI to dynamically choose which backend systems to access, arguing this approach lacks the dependability required for live customer interactions without human oversight.

"A runaway AI bill almost never means the model is too small. It means nobody told it where to look," stated Ashu Roy, CEO of eGain. "And once a customer is talking to the AI directly, there is no one left to catch a mistake before it lands. Getting the underlying information right is not a feature. It is the whole job." Both whitepapers are available on eGain.com.