Enterprise AI · July 2026

The enterprise AI data finally exists. Glean is all over it.

Andreessen Horowitz published hard numbers on where enterprise AI is actually landing. One of the three winning categories has a clear leader in our portfolio, and the data raises a question every buyer should ask about models and context.

5 min readIn partnership withGlean

The data

The scoreboard finally exists

For two years the enterprise AI conversation ran on anecdotes. Vendors claimed transformation, skeptics cited an MIT study saying 95 percent of generative AI pilots never reach production, and both sides argued from sentiment because nobody had hard numbers.

Andreessen Horowitz changed that this spring. Their report, “Where Enterprises are Actually Adopting AI,” aggregates revenue and deployment data from the leading enterprise AI companies instead of surveying opinions, and its findings cut sharply against the pilot-graveyard story.

29 percent of the Fortune 500 are now live, paying customers of a leading AI startup.

The shape of that adoption matters more than the headline. Three use cases dominate: coding, support and search. Within enterprise search, a16z names Glean as the primary startup vendor. We work with Glean, and the data backs a decision we made before it existed.

The problem

What search actually solves

Most enterprises share the same problem. Knowledge lives in hundreds of applications, and the answer to a question sits somewhere in Slack, Confluence, Salesforce, Jira, email or a shared drive while the person who needs it spends part of every day hunting for it.

In the SaaS era that was an annoyance. In the AI era it’s a structural blocker, because a model is only as useful as the context it can reach. An assistant that can’t see your company’s knowledge gives generic answers, and one that sees everything without respecting who is allowed to see what is a data breach with a chat interface.

The hard problem is giving AI your company’s context without giving away the keys.

The platform

Why Glean’s approach holds up

Glean’s core asset is a permissions-aware knowledge graph. It connects to more than 250 enterprise applications, indexes their content, maps how data, people and processes relate, and enforces each source system’s permissions on every retrieval. Search, the assistant and Glean Agents all sit on that graph, so every agent works inside the same permissions as the people it serves.

Two design choices matter to buyers. Glean is model-neutral, supporting more than 40 open and proprietary models through its Model Hub, so a better or cheaper model can be swapped in while the graph, permissions and agents stay put. And governance is part of the product: every agent action is authenticated, permission-checked and logged, with Glean Protect extending those controls.

The objection

Why not just use Claude or ChatGPT directly?

We hear this more than any other question in 2026. The frontier assistants are excellent and increasingly connect to enterprise tools on their own, so if your teams already have Claude or ChatGPT Enterprise seats that reach Drive, Slack and email, it’s fair to ask what Glean adds.

The model and the context layer are different decisions.

A frontier assistant connects to the handful of applications each user authorizes, one connection at a time. That works for an individual but not as an enterprise architecture. Glean keeps a persistent, indexed graph across hundreds of systems, enforces source permissions for every user on every retrieval and logs every agent action centrally.

The second issue is coupling. If one model vendor becomes your context layer, your knowledge, agents and workflows are tied to that vendor’s roadmap and pricing. Glean runs the frontier models underneath, Claude included, so when model leadership shifts (and it has several times in eighteen months) you swap the engine without rebuilding the car.

Swap the model. Keep the context layer. Diagram by Ghost.

For most enterprises the answer is Claude, running through Glean.

Proof

What Glean customers report

Where Ghost sits

Why we are bringing this to you now

Most buyers already know Glean. The risk now is making the wrong architectural call around it: defaulting to whatever assistant came bundled with an existing license, or wiring company knowledge into one model vendor and discovering the coupling later.

The a16z data also confirms the filter we apply across our portfolio. The AI deployments producing real returns own a specific workflow, prove value fast and expand from there. We’ve run Glean evaluations with our customers and can walk you through how it answers the hard questions against your environment and stack.

Deciding on your context layer? Talk to Ghost.

We’ll walk you through how Glean answers the hard questions against your environment and stack.

Talk to Ghost info@ghosttech.co