Iceberg Concept

Definition

The visible layer of AI — tools, models, public data, automation — is above the waterline and accessible to everyone. The competitive layer — context, tacit knowledge, human judgment, relationships, organizational memory — is below the waterline and hard to copy.

The Structure

Above the waterline (visible, commoditized):

• Public AI tools (ChatGPT, Claude, Gemini)
• Generic foundation models
• Publicly available data and benchmarks
• Basic automation and workflow templates

These are available to every competitor. Using them well matters, but they are not the source of lasting differentiation.

Below the waterline (hard to copy, high value):

• Context — the specific constraints, history, and meaning behind decisions
• Expertise — domain knowledge built through years of experience
• Tacit knowledge — what experienced people know but have never written down
• Human judgment — the ability to sense what the data does not say
• Relationships — trust, political awareness, client understanding
• Organizational memory — the accumulated learning of how this organization actually works
• Distinct data — proprietary signals competitors cannot access

The line: “The tool may be common. Your context is not.”

Why This Matters Strategically

LLMs operate by predicting patterns from training data. An organization that feeds a generic model generic prompts gets generic output — indistinguishable from what any competitor with the same tool produces.

Competitive advantage therefore depends on what is below the waterline: the quality of context provided to the model, the expertise used to evaluate its output, and the organizational memory that shapes how it is applied.

This is why an LLM-Wiki — a structured knowledge base of an organization’s own expertise, decisions, and context — is a strategic asset, not just a productivity tool. It makes the below-waterline knowledge visible enough to be queried, challenged, taught, and improved over time.

The Practical Implication

Before asking “which AI tool should we buy?”, ask:

• What context do we have that competitors do not?
• What tacit knowledge exists in our best people that has never been written down?
• What organizational memory is at risk of being lost as people leave or retire?
• What distinct data signals do we have access to that could train or inform a model?

The organization that answers these questions well and builds systems around the answers will consistently outperform the one that simply deploys the latest tool.

What to Pay Attention To

• Where high-value decisions depend on context that has not been captured anywhere
• Where institutional knowledge is concentrated in a few people and at risk
• Where AI tools are producing generic output because they are being fed generic prompts
• Where your organization’s real differentiation lies — and whether it is protected

Connections

Context as Differentiator People Process and Culture Value Equation AI Strengths and Human Strengths Future Skills and Metaskills

Sources

• [inferred from workshop teaching — iceberg metaphor applied to AI competitive strategy; consistent with Brynjolfsson & McAfee on tacit knowledge]

Tags: iceberg, tacit knowledge, competitive advantage, context, organizational memory