eGain: Jensen Huang's AI Stack Lacks Critical Knowledge Layer
eGain Corporation argues that Jensen Huang's model for the AI stack omits a crucial knowledge management layer. The company states this gap often leads to failures in enterprise AI applications.
AI solutions provider eGain Corporation has critiqued the five-layer AI stack model presented by Jensen Huang at GTC 2026, asserting that it overlooks a vital knowledge management component. eGain maintains that this omission is the primary reason for failures in enterprise AI applications.
Huang's model, comprising layers for energy, chips, infrastructure, models, and applications, neglects the critical link between AI models and user-facing applications. eGain identifies this gap as the point where enterprise AI breaks down in practice, particularly when integrating advanced models with existing, often fragmented and outdated, company content.
The company contends that the issue is not with the AI models themselves, such as GPT or Gemini, but with the data they process. Poorly managed, conflicting, or obsolete information leads to inaccurate and inconsistent answers, eroding customer trust and increasing compliance risks. According to Gartner, up to 100% of generative AI virtual assistant projects lacking modern knowledge management integration will fail to meet their objectives.
eGain emphasizes that successful enterprise AI necessitates a dedicated knowledge management layer. This layer must ingest, curate, govern, and update company knowledge to create a unified and reliable source of truth. Only then can AI models reliably synthesize accurate and current information to answer customer queries.