What is a company brain?

Summary

• A company brain is an informal term for a shared, governed layer of business context that AI systems consult: what your terms mean, how metrics are calculated, which sources to trust, and how concepts relate.
• It is shorthand, not a Databricks product name. On Databricks, the closest match is Genie Ontology, which the docs call "the unified context layer" for the Genie family of AI experiences. It is in Public Preview.
• Genie Ontology brings together context you define in Unity Catalog (metric views, domains, and Pages) with context Genie automatically extracts from your existing assets and usage.
• Unity Catalog permissions gate what Genie can use, and citation icons in a response show which knowledge sources informed the answer.

What is a company brain?

A company brain is an informal term for a shared layer of business context that people and AI systems can both rely on. It records what your business terms mean, how key metrics are calculated, which data sources are authoritative, and how concepts, metrics, tables, and teams relate. With that context in place, an agent can use a defined concept to help interpret a request rather than relying solely on inference.
Because the term is informal, this page uses it to mean a governed context layer for AI, and shows how Databricks provides one with Genie Ontology.

How does Genie Ontology work as a company brain?

According to the Databricks docs, Genie Ontology is "the unified context layer that gives the Genie family of AI experiences a business-aware map of your organization." Genie One and Genie Code search the same ontology, so context you curate once applies to both surfaces. The Databricks launch announcement also describes Genie Agents as powered by it. Genie Ontology is in Public Preview.
The ontology brings together two categories of context:

Layer	What it holds	Where it comes from
Unity Catalog semantics	Context you define, govern, and certify	Metric views, domains, and Pages
Inferred context	A map of snippets that Genie extracts and maintains	Existing assets and usage, including metric views, dashboards, SQL queries, and Genie Agents

The docs give three example snippet types:

• Metric definitions, such as "An 'active user' is a distinct user, deduplicated across all platforms."
• Authoritative sources, such as "Revenue questions should be answered using the curated Finance Genie Agent."
• Business rules, such as "A 'qualified lead' only counts once a demo is booked."

What should a company brain do well?

What to look for	How Genie Ontology handles it
Governed definitions	Pages (Beta) hold the definition of a business concept, with fields for domain, owner, synonyms, description, related assets, and sources. Metric views separate measure definitions from the fields (also called dimensions) used to group, filter, and aggregate them.
Context learned from usage	Genie automatically extracts and maintains snippets from existing assets and usage.
A way to rank trust	Each snippet gets an authority score based on where it was generated from, how often it is used, and how fresh it is.
Precedence for human definitions	When a Page defines a concept, Genie One prioritizes the Page's definition over context it infers automatically, and cites the Page.
Permission awareness	Snippets are gated by Unity Catalog permissions, so Genie only uses snippets extracted from assets you have permission to see.
Traceable answers	Clicking the citation icons in a response shows which knowledge sources Genie used.

How do you start building a company brain?

Start narrow. The Databricks guide "Operationalizing Genie Ontology in Your Data Stack" (September 1, 2026) describes six layers, from the data foundation (Layer 0) through metadata, a semantic layer, curated assets, and governance to evaluation (Layer 5). It says you do not have to build all six at once. Its practical advice is to pick one domain like Sales and one metric like ARR, then harden the head. The key idea is to model the head, the authoritative core of concepts you define deliberately, and let Genie Ontology infer the tail.

Who can see what a company brain knows?

Genie only uses content that the person asking is authorized to see, so identical questions can produce different answers depending on the user. Treat the human-modeled layer with care too: Page data does not support customer-managed key (CMK) encryption and is stored in plain text that may be replicated globally, so keep sensitive details out of Page names, descriptions, synonyms, and body text.

FAQs

Is a company brain the same as a data dictionary?

No. A data dictionary describes tables and columns. Genie Ontology also holds metric definitions, business rules, and authoritative sources, and it ranks snippets with an authority score.

Can different teams contribute to the ontology?

Yes. Consumers can create and publish Pages, and they can comment on and suggest edits to published Pages they do not own. Domains organize Pages by business area.

What happens when a business definition changes?

Make the correction where the business meaning belongs: in a Page when the problem is a business definition or synonym, or in a metric view when the problem is a governed measure, dimension, relationship, or calculation. The guide also recommends regularly measuring answer quality to see when the ontology needs to adapt.

Is Genie Ontology generally available?

No. The Databricks docs label Genie Ontology Public Preview, and Unity Catalog Pages is in Beta. Check the current docs before planning a rollout.