AI strategy follows the constraints of the business adopting it

AI strategy follows the constraints of the business adopting it

The claim

Organizations choose different AI starting points because their market position, installed systems, data, regulation, and operating obligations shape what they can safely redesign. The same model capability can therefore produce a process improvement in one company and a new product in another.

Evidence and context

E004 draws on Christopher Holland and Anil Kavuri's 2023 comparison of incumbent insurers and new entrants. Their innovation framework covers products, processes, and the value chain. In the cases studied, incumbents generally applied AI to defend and enhance existing strengths, while entrants used it to build new services with distinct features and customer experiences.

Dalton's account of working across several policy systems gives the distinction an operational form. Historical data can be valuable and difficult to normalize at the same time. A use case that appears simple at the interface may require months of mapping, governance, and reconciliation underneath it.

An entrant without that installed base can design a cleaner path around a narrower problem. It gives up the incumbent's history, distribution, and scale in exchange for fewer dependencies at the start.

The strongest counterpoint

Incumbent and entrant are not permanent technical identities. An incumbent can create a separately governed product with new systems, and a growing entrant can accumulate acquisitions, exceptions, and customer promises that make later change difficult.

The paper itself includes cases rather than a law of organizational behavior. Leadership, architecture, capital, partnerships, regulation, and execution can matter more than company age.

Where the claim holds

The principle is useful when evaluating AI in established, data-intensive, regulated businesses. It helps explain why a firm may begin with augmentation, fraud detection, document handling, or migration rather than a visible customer-facing reinvention.

It should not be used to excuse indefinite modernization, assume every startup is innovative, or predict which business will succeed. A strategy still has to produce customer value and sustainable economics.

Why it matters

AI recommendations become more useful when they begin with the company's actual constraints. "What can the model do?" is only the capability question. Strategy also asks what the business can integrate, govern, review, and support without breaking the obligations that keep it operating.

Where this could lead

This thesis now anchors two source-bounded public drafts: [[Episodes/E004 - Gemini 1.5 Sora and AI Strategy Constraints/Public Drafts/AI Strategy Follows Business Constraints|AI Strategy Follows Business Constraints]] and [[Episodes/E004 - Gemini 1.5 Sora and AI Strategy Constraints/Public Drafts/Compare Incumbent Insurer and Insurtech AI Strategy|Compare Incumbent Insurer and Insurtech AI Strategy]]. Future revisions should add current carrier and insurtech cases, regulatory evidence, implementation outcomes, and economics. They should also test when a separately governed business or partnership lets an incumbent escape its usual constraints.