Enterprise AI Materiality Matrix

Baron Haussmann, appointed Prefect of the Seine in 1853, initiated his transformation of Paris with two strategic axes, one east-west, the other north-south, crossing at Châtelet. This simple grande croisée helped change Paris’s circulation logic and began reshaping a city made of thousands of streets.

The same approach ought to apply to the integration of AI into corporate operating systems. The question is no longer whether companies have AI use cases, since most do. It is whether they are working on those that can fundamentally change the economics of their business, i.e., whether they have found their grande croisée (see also ‘10 Thoughts About Corporate AI’, 2025).

The prize is big. As recently argued in ‘Applied Intelligence’, ‘AI winners’ will emerge in every sector to establish themselves as industry leaders, with superior financial profiles and premium valuation levels.

With this in mind, I worked with Claude to build an indicative ‘Enterprise AI Materiality Map’ inspired by the SASB Sustainability Map. Here is the output, with darker cells indicating areas where AI is more likely to be financially material, not merely operationally useful:*

At its simplest level, AI materiality follows the economics of the sector: operations in asset-heavy sectors, sales and marketing in consumer-facing sectors, innovation in R&D-intensive sectors.

A materiality map is as relevant to executive teams as it is to investors and M&A practitioners. AI-enabled mining of corporate disclosure (annual reports, sustainability reports, press releases, corporate websites, job postings, patent filings, …) can already help benchmark a company’s AI maturity across its peer group, with a focus on material sources of value creation. In the coming quarters, firms should expect rising scrutiny from investors in earnings calls, roadshows, investor days, and, on the M&A front, management presentations and due diligence sessions.

Extracting the value of enterprise AI is particularly relevant for firms serving the supply side of AI. If they can tangibly raise productivity in their own operations through AI, the message to the market becomes more coherent: the infrastructure they help build is not speculative plumbing, but the backbone of a broadening enterprise adoption cycle.

As enterprise AI moves from experimentation to measurable productivity, the data center debate will evolve further. The investment case will no longer rest primarily on hyperscalers’ capex plans (supply), but increasingly on a broader corporate adoption cycle (demand).

This evolution may provide a new source of support for the valuation of firms supplying goods and services to build the AI infrastructure. It may also provide another kicker to financial markets globally.

* Sources: Claude Opus 4.6, BCG ‘Build for the Future’ (2025), McKinsey ‘Economic Potential of Generative AI’ (2023), McKinsey ‘State of AI’ (2025), sector-specific research.