Enterprise AI’s next battle is over workflows, data and distribution

Hello,

With AI moving deeper into the workplace, developer habits and proprietary data are becoming as important as the models themselves.

As companies embed AI into everyday work, the bigger advantage may come from controlling developer habits, proprietary data and the infrastructure connecting businesses to AI models.

The Information is reporting from inside that shift. Subscribe for $299 and save 25% on your first year for reporting that gets inside the companies and strategies reshaping the technology industry—before it becomes mainstream.

Developer habits become a moat

OpenAI is discounting some of its top models by 50% on OpenRouter as it tries to gain ground on Anthropic among developers and strengthen its business-to-business push.

That fight comes as Anthropic’s Claude Code keeps gaining enterprise users, even as its cost per user has climbed sharply. OpenAI’s Codex is growing quickly too, but access alone hasn’t been enough to dislodge developer preferences.

The next data gold rush

As AI labs exhaust more of the public internet, they are increasingly looking for something harder to find: data showing how people actually work.

Startups are fielding offers for Slack messages, GitHub activity, meeting transcripts and other internal data, giving AI developers a window into real workplace behavior—including mistakes.

That could make proprietary enterprise data one of AI’s most valuable new assets.

The battle for the gateway

Companies increasingly want access to multiple AI models rather than betting on one provider.

That has fueled demand for routing platforms such as OpenRouter, while Cursor, Databricks, Palantir and others build their own systems to decide which model handles each task.

The prize is significant: control the routing layer and you may control how enterprises buy and use AI.

AI rewrites software economics

The shift is already changing software businesses.

Sierra is moving beyond customer support, AlphaSense is changing how it charges customers as AI agents account for more activity and Microsoft is using AI to overhaul how its security organization operates.

At the same time, specialized AI apps are showing the model labs won’t necessarily win every market. A fast-growing AI product for doctors has kept expanding even as OpenAI enters healthcare.

The emerging lesson: the model is only one part of the product. The real enterprise AI battle is over the workflows, data and distribution surrounding it.

The Information’s subject specialists are already inside the companies and conversations shaping that shift. Subscribe now and save 25% to follow the exclusive reporting that goes beyond the headlines.

Our reporting is now easier to access wherever you are with the new The Information app, available on the App Store and Google Play.