Back to all articlesWho Owns the AI Context Layer? A Control Tower for Data Teams
AI

Who Owns the AI Context Layer? A Control Tower for Data Teams

The data team should own how an AI analytics agent reaches an answer. Business experts should own what the business rules mean. Together, they need a correction process that identifies the source of a mistake, names who can change it, and checks the next answer.

Imagine a revenue investigation that uses the approved calculation but cites an expired promotion as the cause of a decline. Finance may have nothing to change in the metric. The promotion's owner needs to correct the supporting context, and the data team needs to verify where that correction takes effect. That is the job of a context-layer control tower.

Orion by Gravity provides documented building blocks for this approach, including editable knowledge pages, change requests on locked pages, reusable analysis notebooks, and MCP access from other assistants. The control-tower workflow below is a recommended operating model; its complete implementation depends on the tools and integrations you choose.

Product and source documentation reviewed October 16, 2026. The revenue scenario is hypothetical. The GitHub example is a first-person Gravity account, checked against the correction and audit messages and a saved metric definition and result history. It is one internal workflow, not a measured accuracy study.

What is an AI context layer, and why does it need a control tower?

An AI context layer is the maintained information and rules an agent uses to interpret a question and work with your business data. It can include metric references, business terminology, operating events, procedures, approved analyses, and rules about which context applies to which audience.

The term is still developing. Define its responsibilities before buying a product or creating another service in your stack.

A knowledge repository gives the agent something to read. A control tower gives people a way to manage the consequences: who can change that information, which agents consume it, what answers it affects, and how a mistake gets resolved. A glass-box experience exposes inspectable evidence such as queries, source documents, and selected definitions. It does not require access to a model's private reasoning.

For example, imagine that a sales user asks why revenue fell. The agent returns a valid query and a persuasive explanation. The query might use the wrong revenue definition, or the explanation might rely on an outdated note about a product launch. Those failures need different owners and different fixes.

This is why context ownership should be an operating responsibility. An upload screen alone cannot tell you whether the business answer changed for the right reason.

Does a context layer replace the dbt semantic layer?

Keep governed metric calculations in their authoritative system. The dbt Semantic Layer centralizes metric definitions on existing models and uses MetricFlow to handle joins. A context layer can add the business explanations, scope, and operating knowledge needed around those definitions.

That boundary is practical, rather than a fixed divide between structured and unstructured data. In its September 16, 2026 context-engineering article, dbt describes bringing versioning, provenance, and tests to the text agents read. Some context work may therefore belong in your dbt project too.

Choose one source of truth for each rule. If Finance changes net revenue, review the calculation through your metric-governance process. If Sales adds an explanation of a one-time promotion, review the business note with its owner. Link the note to the metric; avoid maintaining a competing revenue formula in prose.

For a fuller explanation of this boundary, see Semantic Layer vs. Business Context.

What should a data team's context-layer cockpit let it control?

Use these six controls as an acceptance checklist. They are design requirements to verify in a pilot, rather than a claim that every context product includes them.

ControlQuestion the technical team should be able to answerEvidence to request
ScopeWhich user, project, customer, and tool does this rule apply to?Effective identity, access configuration, and selected context
MeaningWhich approved definition should the agent use?Reference to the authoritative metric plus owned business notes
ChangesWho proposed and approved this correction?Proposed difference, reviewer decision, and effective date
AnswersWhat information and executable logic produced this result?Query or notebook, source references, and returned result
QualityDid the change fix the intended answer and preserve other answers?Before-and-after checks against a maintained question set
DistributionWhich assistants actually received the correction?Integration path, refresh behavior, and a check in each consuming tool

Start with one important business question and exercise all six controls. A system that can show SQL but cannot explain which definition applied leaves the team with an incomplete investigation. A system that stores approved context but has no adoption path into the consuming agent leaves the correction unused.

The cockpit should connect an incident to a decision: keep the answer available, request clarification, route it for review, or temporarily limit the affected experience. Decide which actions you can support before promising them to business users.

Who should own definitions, approvals, and the business-user experience?

Assign ownership by the kind of decision involved.

  • Business experts own meaning. Finance owns revenue policy; Customer Success owns the interpretation of a customer-health playbook. Each entry should identify a responsible owner and when it applies.
  • Data teams own the path to an answer. They maintain metric references, context selection, test questions, and the evidence needed to investigate a result.
  • Administrators own permissions. They decide who can read, propose, approve, and change configuration. Editing business meaning and granting data access should be separate decisions.
  • The product or experience owner owns the user journey. They decide when users can explore freely, when a request needs clarification, and when a reviewed answer is appropriate.

These responsibilities can belong to a small team. They still need explicit boundaries. Giving someone permission to correct a business note should not silently grant permission to redefine a governed metric or expand warehouse access.

What does a human correction look like in practice?

In Drew Gillson's account of repairing Gravity's GitHub data pipeline, a new source produced higher commit counts than the old source. Gillson suggested that the earlier metrics had counted commits on main branches. Orion tested that explanation and reported that the filter reduced the differences, while some gaps remained. Gillson approved the proposed change after seeing those limits.

A follow-up audit found that related leaderboards still counted a different set of branches and that some metrics included a partial day. Gillson chose a shared branch rule and complete-day windows. The saved metric definition and result history inspected for this guide show the branch filter and the later complete-day calculation; the project documentation records that choice.

The useful lesson is to preserve the question, test the owner's interpretation, disclose what remains unexplained, and check other consumers of the same definition. A correction can improve consistency while still leaving a reconciliation task open.

Read How I used Orion to repair our GitHub data pipeline for the first-person account. This example shows a person directing and approving changes in one internal workflow. It does not establish mandatory independent approval, automatic correction of every consumer, or control over external agents.

Can business experts update agent knowledge without changing code?

Yes, for knowledge maintained as editable business content. That does not make every change a no-code change: altering a metric calculation, warehouse policy, or integration can still require engineering work.

Orion by Gravity documents page editing, revision history and change requests. The page-management guide says admins can edit locked pages directly and creators can bypass review for their own edits. The MCP update_wiki_page reference describes a narrower direct-edit exception: the creator; other callers submit a change request. Verify the interface and role your team uses. A locked page alone does not establish mandatory independent approval.

The same documentation describes project-selected pages and optional tenant or group defaults. Review scope carefully before changing a default, because it can affect multiple conversations.

These controls let business experts contribute corrections without asking an engineer to rewrite the agent for every note. The operating process still needs a named owner and a check of the resulting answer.

What should happen when a business user reports a wrong answer?

Assign an incident owner before changing shared knowledge. That owner should preserve the disputed question and available evidence, identify whether the problem belongs in data, a governed calculation or a business note, and send it to the owner of that source. The reviewer decides which interpretation is justified and who should receive it. The data team checks the answer through the intended consuming path.

Keep the incident pending until that check is complete. A saved edit proves that content changed; it does not prove that the next answer used it. Orion's analysis notebooks retain instructions, logic and code and can be rerun, providing an artifact to inspect. Verify separately which context revisions and historical inputs are available. For the submission and review steps, use How Business Users Can Correct AI Agent Context Without Changing Code.

Who decides whether a correction is ready?

The business owner approves meaning; the data owner checks its application. Agree on one question that should change and one boundary that should stay unchanged. A more consistent answer is not necessarily a correct one: check the metric, reporting period and supporting evidence, not just whether repeated wording matches.

Use How to Test an AI Context Layer After dbt or Business Rules Change for the test cases and comparison method. Keep data tests in dbt and agent-answer checks alongside them. Where your tools lack automated gates, record a manual review and a recovery decision. These are operating recommendations; they do not establish a native Orion CI gate or automatic rollback.

How do we keep Claude, Slack, and BI tools using the same context?

Consistency requires a verified consumption path. A shared page cannot govern an assistant that never reads it, and a connected assistant may still have access to other tools that bypass your intended path.

Inventory the entry points. For each one, document how it retrieves the authoritative metric, which business pages it can read, whose permissions apply, and how an update becomes effective. Test the same question in each surface. If a tool cannot consume the shared context, make that limitation visible and assign an owner to its separate configuration.

Orion's MCP server gives compatible assistants, including Claude, access to Orion projects, metrics, workflows, and knowledge pages. This is one documented path for using Orion from another assistant. It does not, by itself, make every Claude session, Slack bot, or BI tool use the same definitions, or give Orion control over their independent execution.

Treat cross-agent oversight as an integration requirement to demonstrate in your environment. Ask to see an actual request, its effective identity, the selected context, the answer artifact, and the effect of a correction.

Should we build a context-layer control tower or buy one?

Build when your team needs custom execution paths and can maintain context selection, access enforcement, change review, integrations, and evaluations over time. Buying can provide existing editing and analysis surfaces, but you still need to verify that the product supports your owners, metric systems, and consuming agents.

Check what your existing stack already covers. dbt governs metric calculations. Hex's customer accounts describe context maintained in existing repositories and reviewed through team workflows. Supper describes shared definitions, activity visibility and access controls in its Admin Control Center. DataHub describes a workspace where subject-matter experts refine context and resolve definitions. A shared knowledge platform may cover document maintenance. Compare the missing responsibility in your own correction process before buying another tool.

Use a pilot to follow one correction from report to verified answer. Include a scope boundary and a change that should leave another audience unaffected. Measure the work your team must do to complete the loop; evaluate capabilities through demonstrated behavior.

Frequently asked questions

Is a control tower the same as an AI agent observability dashboard?

It can include observability, but the operating model also needs ownership, change review, context scope, and a correction path. An answer trace helps explain an incident. The control tower connects that evidence to a reviewed action.

Does a context layer prevent hallucinations?

No layer guarantees correct answers. Approved sources and governed definitions make some errors easier to prevent and investigate. You still need checks of source selection, calculations, access, and unsupported explanations, plus a clear response when evidence is insufficient.

Do locked Orion knowledge pages require approval for every edit?

No. The page-management guide describes admin and creator exceptions; the MCP update tool describes direct edits by the creator. Confirm the actual role and interface in your deployment before relying on a locked page as an approval requirement.

Does Orion's dbt connection guarantee that every answer runs through MetricFlow?

Do not assume that. The connection guide documents dbt context such as model descriptions and lineage. Confirm the metric execution path separately in your deployment; importing dbt metadata is different from enforcing Semantic Layer API usage.

Where should we start if we have several data agents already?

Choose one important question used in two surfaces. Map its metric source, business context, owners, permissions, and correction process. Demonstrate a reviewed change in both surfaces before extending the pattern to additional agents.

What should we ask Orion by Gravity to demonstrate?

Bring Orion by Gravity one disputed answer and the source your team considers authoritative. Ask to see an ordinary contributor propose a correction, the reviewer decision, the resulting source change, and a checked answer through the tool your users actually use. Then repeat with the page creator and an admin to identify review exceptions. Record what the demonstration shows and which steps require integration or a manual process.

Keep reading

View all