A Practical Guide · Version 3 · October 2026

Your AI Stack, Governed

Your organization already owns the most valuable layer in AI: the context that defines what your data means. This guide shows how Genie One becomes the front door for governed data questions, how Genie Code builds the assets behind it, and how MCP delivers those answers into Teams, Microsoft 365, and the AI tools your teams already use.

Covers Genie One, Genie Code, Unity Catalog, and MCP integration with Microsoft 365 and third-party AI assistants
One AI coworker, everywhere your teams work
💼 Executives 📈 Business teams 🔍 Analysts 💬 Microsoft Teams 💻 Microsoft Copilot 🤖 Claude ✨ OpenAI 📊 Microsoft Excel 💬 Slack 📋 Google Sheets
⚡

Genie One

The data-smart AI coworker — ask, get answers, take action
“What is our same-store NOI this quarter?”
↓
🧠 Genie Ontology A self-improving knowledge graph of your data, documents, and definitions — shared by every agent and surface
🔒 Unity Catalog One copy of the data, one access model, one place where definitions live

Executive Summary

⚡
Start here

Genie One is your data-smart AI coworker

Not just a place to read answers — a place to ask questions, get answers grounded in your business context, and take action on what you learn. Business users work in plain language while Genie One handles the data, the documents, and the follow-through, all within your existing Unity Catalog permissions.

Native to Databricks No SQL or notebooks Web, desktop, mobile Slack & Teams Excel & Google Sheets MCP app · REST API coming Runs on your Unity Catalog permissions
⚡

Real AI-driven outcomes

Ask, analyze, then act. Schedule recurring tasks, set alerts, save a workflow as a reusable skill, and turn any conversation into a polished document to share by link, email, or PDF.

🧠

Understands your business context

A self-improving knowledge graph built on your data, documents, and definitions — so Genie knows your business as well as your teams do, and gets better as they use it.

🔒

Unified and secure governance

One permission and audit model spanning data, documents, and every connected tool, governed by Unity Catalog and Unity Gateway.

Behind the coworker: anyone can turn a Genie One conversation into a Genie Agent scoped to one use case, with its own instructions, skills, and sources. Genie Code is where analysts and engineers deepen the governed context over time. MCP carries that same context into Slack, Teams, Microsoft 365, Claude, and your own applications. Unity Gateway governs and meters all of it.

The principle underneath all of it: invest once in the context layer unique to your organization, let every AI surface benefit from it, and govern and measure the whole estate in one place.
1

Every AI tool is a stack of two layers

Separating them clarifies where your investment actually compounds.

The Intelligence Layer

What you rent

The reasoning, generation, and conversational capability of a frontier model. This is the primary product of Claude, Copilot, and most general-purpose AI assistants.

  • Improves rapidly, with new model releases every few months
  • Increasingly commoditized across vendors
  • Typically re-evaluated every 12–18 months
  • Low switching cost, and therefore no durable advantage
Valuable capability — but paying premium rates here to compensate for missing data context is an expensive substitution.
The Context Layer

What you own

Your tables, permissions, lineage, metric definitions, semantic rules, and curated Genie Agents — governed in Unity Catalog and attached to real data assets.

  • Unique to your organization and your business logic
  • Compounds over time as definitions and trusted assets accumulate
  • Cannot be purchased from or replicated by a vendor
  • The durable advantage: it is what makes an answer yours
Investment here carries forward. Every AI surface you connect to it inherits the same governed understanding.

Genie One and Genie Code are built to operate directly on the context layer.
Claude and Copilot remain useful for their own strengths — and through MCP they can draw on the same governed definitions rather than relying on context pasted into each conversation.

One question, five paths to an answer

An illustrative scenario: “What is our same-store NOI this quarter?”

Approach Likely outcome Why Result
💻Copilot
Surfaces 118% from a board deck or spreadsheet circulated in a previous quarter. The retrieval target is a document rather than the governed dataset. The figure may have been accurate when written, but the document has no refresh cycle. Point-in-time
🤖Claude
no data connection
Asks you to supply the data and clarify how your organization defines same-store NOI. No route to the warehouse, and no visibility into your approved metric logic, exclusions, or cohort policy. Needs input
🤖Claude
connected via MCP
Calls the Genie MCP endpoint, which runs your approved logic and returns +3.4%. The model itself is unchanged. What changed is that it now draws on the governed context layer you maintain. Governed
⚡Genie One
Returns +3.4% along with the SQL behind it, a segment breakdown, and the reviewed dashboard that already tracks the metric. Purpose-built for this pattern. It searches existing governed assets first, applies the curated Agent definition, and runs read-only SQL within Unity Catalog permissions. Governed + context
🛠️Genie Code
Reaches the same answer, then builds or refines the dashboard, adjusts the metric logic, and publishes a reusable asset for the team. The analysis becomes a governed artifact. Next quarter, Genie One surfaces it directly rather than reconstructing the work. Reusable asset

What actually changes between rows two and three

The model is identical in both cases. The intelligence did not improve — the context layer beneath it did. This is the practical case for investing in governed context: it raises the quality ceiling for every AI surface you connect, including the ones you have already purchased.

Outcomes above are illustrative and depend on how each tool is configured in your environment, what data it can reach, and whether your metric definitions have been documented.

How one question picks up everything it needs

Row three above is only possible because something sits underneath it. This is that something — the same question, followed down through the layer, picking up at each step what it did not arrive with.

❓ “Which assets are tracking below underwritten NOI, and what explains the variance?” Asked once
1
Connect
Delta LakeIcebergSnowflakeBigQuerySAPTeradataSalesforceSQL ServerWorkspace assetsBusiness apps
Where the answer could live — federated data stays where it is, rather than being copied
2
Govern
Unity CatalogUnity Gateway
What this particular person may see — fund and joint-venture segregation holds before a single row is read
3
Context
Learned knowledgeOntoRank · authority, relevance, permissionsModeled knowledge Genie Ontology — continuously learned context for enterprise AI
What the words actually mean — which NOI definition, which underwriting version, which period — and which source carries the authority
4
Consume
Genie OneGenie AgentsGenie CodeAgent BricksClaude, ChatGPT, Copilot, your app · via MCP
Answered in whichever surface asked — with the figure, the SQL behind it, and citations to the sources used
↻
Improve
Proactive capture · private to youAutomatic inference · inferred, not certifiedDeliberate authoring · shared governed asset
What the layer keeps for next time — so the next question starts further along than this one did

Open connectivity to any catalog, database, app, or MCP — 50+ connectors. Three paths into the layer, three different scopes: saving something for yourself is not the same as publishing a company definition, and inferred context is not certified context.

What actually happens at the Context step

Behind the scenes · in milliseconds
1A question is asked in Genie — or from another agent through Genie MCP
2Genie retrieves and ranks the relevant learned and modeled knowledge, while respecting that user’s permissions
3The selected knowledge is injected into the reasoning loop before the model answers
4The answer comes back grounded in enterprise context, with its sources cited
👍👎User feedback on answers continuously improves the ranking — which is why the layer gets sharper with use rather than staler.
One question, two kinds of knowledge — resolved together
“Which assets are tracking below underwritten NOI, and what explains the variance?”
Modeled knowledge · defined Same-store NOI Domain: Asset Management Rental revenue less operating expenses, for assets held the full comparable period. Codified deliberately as a governed definition. Metric view · noi_actuals
Learned knowledge · inferred Underwritten NOI Inferred from: acquisition model notebook The NOI projected at acquisition. Nobody wrote this definition down for Genie — it was discovered from the artifacts your team already produced. Metric view · underwriting_targets

OntoRank scores both by authority, relevance, and the asker’s permissions, then passes only what the question needs. One term was defined on purpose; the other was learned from work already done. Neither reaches the model unless this person is entitled to it.

What you own afterward

Every answer leaves something behind

The architecture above is a snapshot. The reason to build here is what accumulates inside it, because the durable asset is the business meaning — not the last answer.

+1Certified definitionSame-store NOI against underwriting, agreed once by the asset manager accountable for it, then applied everywhere after.
+1Trusted query or dashboardThe validated logic behind the answer, still standing next quarter when the question returns.
+1Curated Genie AgentA scoped agent that now carries this domain and its rules for whoever asks next.
Each one raises the floor for every surface above it. The next person to ask — in Genie One, in Excel, or through Claude over MCP — starts from the asset instead of from scratch. That is what compounds, and it is why this layer is worth owning rather than renting.
2

Three surfaces, one reinforcing cycle

How Genie One, Genie Code, and MCP work together rather than overlapping.

⚡

Genie One

Ask, answer, act
Most of your organization asks questions in plain language, then acts — scheduling tasks, setting alerts, saving skills, and producing documents.

→reveals demand
🛠️

Genie Code

Deepen the context
Analysts and engineers enrich the ontology — metric views, certified assets, pipelines — while any user can turn a conversation into a Genie Agent.

→publish & expose
🔌

MCP

Distribute
The same governed context reaches Slack, Teams, Microsoft 365, Google Sheets, Claude, OpenAI, and your own applications.

Why the cycle matters. Genie One reveals which questions people ask repeatedly and where answers fall short. That feedback deepens the ontology, and MCP carries the improved context into every connected surface. Because the knowledge graph is self-improving, each pass raises quality and lowers effort — the next person to ask gets the better answer without anyone rebuilding it.

Matching people to the right surface

Most organizations need far more consumers than creators.

Group Primary surface Usage pattern Why it fits
Executives and senior leaders Genie One, or Microsoft Teams via MCP Occasional, high-stakes They need trusted answers, not an authoring environment. Delivering into Teams removes the need to change tools.
Business teams and managers Genie One Regular question-asking Governed exploration without needing to understand compute, SQL, or notebooks.
Analysts and power users Genie One and Genie Code Both consume and create They ask questions and also build the governed assets that answer them next time.
Data engineers and platform teams Genie Code Sustained authoring They own pipelines, definitions, shared assets, and the quality of the governed layer.
Existing AI assistants Connected via MCP Complementary Claude and Copilot keep their own strengths while drawing governed figures from a single source.

You are not starting from scratch

Genie learns from work your teams have already done. Certify your existing high-quality Genie Agents, dashboards, queries, and notebooks, and enable ontology snippets — Genie draws on them directly, provided your users have access to those assets so the relevant snippets can surface. Genie Code is how you deepen and maintain that context over time, not a prerequisite for turning Genie One on.

How this fits alongside your existing BI investment

Genie One is a discovery and conversation layer over the assets you already have — it searches your published dashboards and queries before generating anything new. Tools such as Power BI remain valuable for standardized reporting; connecting them to the same Unity Catalog data means the figure in a report and the figure in a Genie answer reconcile to the same source.

3

Genie Ontology is the durable advantage

A self-improving knowledge graph that connects your sources into one shared understanding of the business.

📚

What the ontology contains

Far more than a glossary. It connects approved datasets, metric views, semantic rules, join logic, filters, fiscal calendars, exclusions, certified assets, and learned snippets into one graph — working off a shared vocabulary so every agent and every surface resolves a term the same way.

🧠

Why general-purpose AI struggles here

Revenue, occupancy, yield on cost, and same-store NOI are ordinary phrases with organization-specific meanings. No model can infer your Finance policy from vocabulary alone.

🔒

Why it compounds rather than decays

Definitions live beside the governed data, permissions, lineage, and audit trail in Unity Catalog — and the graph improves as teams use it, certify assets, and correct answers. It is a managed system of record for meaning, not a collection of prompt conventions.

How your knowledge reaches Genie

Connectors and MCP servers expose context where it already lives, rather than copying it into prompts.

📄
Documentation systems Runbooks and operational docs stay where your teams maintain them. Exposed once, then drawn on whenever a question actually needs them.
ConfluenceSharePointGoogle DriveAI Search index
💬
Business applications Native connectors for the tools your teams already work in — so a decision recorded in a thread or a ticket counts as context too.
SlackTeamsAtlassianGmailCalendarGitHub
🗃️
Data sources Data already in Databricks is reached through a Genie Agent in plain language. Data held outside it connects through an external MCP server.
Unity CatalogGenie AgentsExternal MCP
⚙️
Tools and custom apps Platform APIs, validation and deployment scripts, and proprietary workflows become things Genie can do, not merely read about.
Internal APIsDatabricks AppsCustom MCP
🔒 Administrators decide what is reachable. Workspace admins control which connectors and MCP servers are available, and users select only from those approved sources. Every response respects the permissions of the source it drew on — so connecting more context never means exposing more of it. Over 50 connectors are available today.

Context is stored as snippets, not documents

Six kinds, each carrying the permissions of wherever it came from.

🧮
ExpressionsThe formula, exclusions, and time window behind a measure
🔗
RelationshipsWhich join path between two entities is the approved one
🏷️
TopicsThe subject area a question belongs to
👤
ExpertsWho authored a definition and still maintains it
⚙️
LogicFiscal calendar, qualifying criteria, what counts and what does not
🔒
PermissionsAccess constraints carried from the originating source
OntoRank
Three definitions of “same-store NOI” exist across your estate. OntoRank decides which one answers the question.
Asset management dashboardReviewed · opened daily · 40+ viewers
● Selected
Certified metric viewPublished · moderate use
Candidate
Draft notebookSingle author · untouched 8 months
Down-ranked
Authority is inferred, not declared — weighted by how an asset is used and who authored it, then filtered by the asker’s permissions before anything reaches the model. Relative weights shown are illustrative.
👁️
ReviewSee what the ontology has extracted
✅
AuditCheck what is being treated as authoritative
📤
PublishCorrect it, or promote your own definition

Built automatically, but governed deliberately — administrators curate all three through the semantics glossary.

What gets remembered, and for whom? Personal memory versus shared governed context
Personal Remember this for me
“I prefer revenue analysis in USD.”
1You mention something useful in conversation
2Genie One proposes remembering it
3You approve, modify, or reject the suggestion
4If saved, it applies in your later conversations, cited when used
ScopePrivate to you
ApprovalYours to approve, modify, or reject
ReuseYour later Genie One conversations
ControlAsk Genie to update or delete it
Shared and governed Define this for the business
“Approved revenue excludes assets under major renovation and one-time items.”
Team-defined semanticsMetric views, Pages, domains — explicitly governed and certified
Inferred contextSnippets drawn from supported dashboards, queries, metric views, and Genie Agents
Both feed Genie Ontology — ranked by source authority, usage, and freshness, retrieved inside your Unity Catalog permissions, and able to cite the sources it used.
Changing a shared definition has its own path Genie Code helps draft it, the business or data owner validates the logic, and an authorized person publishes the governed asset. That gate applies to proposed changes — not to every inferred snippet.
Genie OneBusiness questions
Genie CodeTechnical analysis and asset creation
Approved assistantsGoverned answers through Genie One MCP, Claude included
Shared does not mean visible to everyone — retrieval stays scoped to each person’s permissions on the underlying data.
⚠️ These are two paths, not one pipeline. Saving a personal memory does not publish an organizational definition. Promotion is a deliberate workflow: propose the rule, have the owner validate it, then publish the appropriate governed asset. A personal preference need not become a company rule — and a validated company rule should not depend on every user remembering to paste it into a prompt.

How this compares to Claude

Claude also supports persistent context and reusable outputs — projects, instructions, files, and memory. So the useful question is not whether an assistant can remember. It is where the meaning lives, who is entitled to reuse it, which governance applies, and how it becomes an organizational asset. That is where a governed context layer tied to your data differs from context held inside a conversation.

Examples above are illustrative. Genie Ontology is documented as Public Preview — confirm availability and configuration for your own environment.

What a Genie Agent adds on top of the ontology

A conceptual view of how ambiguous terms become consistent answers.

# Conceptual structure for an asset performance agent { "datasets": [ "main.asset_mgmt_gold.noi_actuals", "main.asset_mgmt_gold.underwriting_targets", "main.asset_mgmt_gold.property_operating_expense" ], "business_definitions": { "same_store_noi": "Rental revenue less operating expenses for assets held the full comparable period, excluding management fees", "reporting_period": "Fiscal quarters begin February 1", "approved_filters": "Exclude assets acquired mid-period, dispositions, and assets under major renovation" }, "join_rules": [ "property_operating_expense.property_id = noi_actuals.property_id", "underwriting_targets.month_start = noi_actuals.month_start" ], "response_rules": [ "Always state the reporting period", "Indicate whether the figure is actuals, plan, or forecast", "Return the SQL used to produce the answer" ], "trusted_assets": [ "Asset Management NOI Dashboard", "Quarterly NOI Variance Query" ] }
Illustrative structure rather than exact configuration syntax. The principle holds regardless of interface: approved data, approved logic, explicit response rules, and links to assets that have already been reviewed.
Without a governed layer Metric logic lives with individual analysts. Everyone else asks in chat, reads older decks, or copies last quarter’s SQL.
→
With a governed layer Definitions are encoded once, and every connected surface inherits them automatically.
Definitions supplied per prompt An assistant only applies your rule if the user remembers to include it in that particular conversation.
→
Definitions retrieved automatically The approved definition is fetched from the governed layer, removing the reliance on individual recall.
Figures diverge across tools One number in a spreadsheet, another in a slide, a third in an AI conversation.
→
One definition, many surfaces Genie One, Teams, and connected assistants reconcile because they query the same governed source.

How Genie Agents relate to the ontology

The ontology is the shared layer. A Genie Agent is a scoped view of it — a conversation narrowed to one use case, with its own instructions, skills, and sources, deployable in Genie One, Slack, Teams, or a custom application. Anyone can create one by turning a chat into an agent. Where a domain needs tighter oversight, a subject-matter expert curates the agent, and its context feeds back into the shared ontology so everyone else’s answers improve too.

In one sentence

Models are rented and replaced. The ontology describing your business is owned and compounds — which makes it the part of the stack most worth building deliberately.

4

Delivering answers where people already work

Adoption improves when the governed answer appears in the tool that is already open.

⚡
Genie One
The front door for governed data questions, and for the work around them — drafting shareable documents and searching connected sources such as SharePoint and Google Drive.
Native
💬
Microsoft Teams
An out-of-the-box Teams app. @mention Genie in any channel and it replies in the thread, or direct-message it to work with your enterprise context from anywhere.
Native app
💻
Microsoft 365 Copilot
Strong for documents, email, meetings, and drafting. MCP lets those workflows source governed figures from Genie rather than from circulated files.
Configuration required
🤖
Claude and OpenAI
Point an MCP client such as Claude or ChatGPT at the Genie One MCP server and authenticate with OAuth. Governed answers then come from Unity Catalog rather than pasted context.
Configuration required
📊
Power BI and Excel
Keep the reporting surfaces your teams know, pointed at governed Databricks data so figures reconcile with Genie answers.
Supported
📦
Databricks Apps
Package a purpose-built experience for a specific team workflow while remaining on the same governed backend.
Supported
💬
Slack
An out-of-the-box Slack app. @mention Genie in a channel for an in-thread reply, direct-message it, or let it post on your behalf through the chat app MCP.
Native app
📋
Google Sheets
Ask Genie from the Databricks Connector sidebar and work with governed answers, and their citations, without leaving the sheet.
Supported
💻
Desktop and mobile
A macOS desktop application plus native iOS and Android apps, so teams in the field keep the same governed context they have at their desk.
Native
⚙️
REST API
Programmatic access for embedding governed Genie answers into your own services and workflows.
Coming

What this looks like in practice

The same governed figure — +3.4% — delivered across four different surfaces your teams already use.

Microsoft Teams # finance-leadership
JD
What is our same-store NOI this quarter?
⚡
+3.4%
Same-store NOI for Q1, using the approved definition and excluding assets under major renovation.
via Genie One Unity Catalog permissions applied Source: Asset Management NOI Dashboard
Microsoft 365 Copilot Board update.docx
JD
Draft the performance paragraph for the board update using our current same-store NOI.
✦
“Same-store NOI closed the quarter at +3.4%, an improvement on the prior period, driven primarily by the multifamily portfolio segment.”
figure retrieved via Genie One MCP governed source Drafting by Copilot
Genie One · Mobile
JD
NOI trend?
⚡
+3.4%
Up from +2.1% last quarter.
Genie One governed
Microsoft Excel Q1 Board Pack.xlsx
ABC
1PortfolioSS NOINOI Δ
2Multifamily+4.6%+$2.4M
3Student housing+3.1%+$310K
4Hospitality−1.2%−$84K
5Total+3.4%+$2.6M
⚡ Genie One
Break same-store NOI down by portfolio
Written to A1:C5 using the approved definition.
via Databricks Add-in UC permissions applied

Illustrative representations of the interaction pattern across four surfaces, not product screenshots. Each surface renders results in its own native style, and every one enforces the same Unity Catalog permissions.

What “governed” means in practice

The same three guarantees hold on every surface above, whether the question starts in Genie One or arrives through MCP.

📐

One definition, applied everywhere

Metric logic is encoded once in the governed layer rather than restated in each conversation. Whoever asks, and wherever they ask from, the same approved definition, filters, and fiscal calendar apply.

🔍

Every answer is traceable

The generated SQL, the lineage behind it, and the access boundaries that applied are all inspectable. You can show not only the figure but exactly how it was produced.

🔐

One permission and audit model

Every request is authorized against your Unity Catalog permissions, whether it reads a governed table or searches SharePoint. Unity Gateway applies policies and approval requirements, and records usage and audit logs in system tables.

People choose the surface. The source of truth stays the same.

Genie One covers more ground than a data tool alone. Alongside governed metrics it drafts shareable documents in a canvas beside the chat, and connects to SharePoint, Outlook, Google Drive, Slack, Jira, and Confluence through Databricks-provided connectors, each authorized against your Unity Catalog permissions. Where teams would rather stay in Microsoft 365 or another assistant, MCP delivers the same governed figure into that surface. Either way, the definition behind the number does not change.

Governance and security considerations

MCP does not bypass governance. It exposes approved tools and governed context through a controlled interface, and Unity Catalog permissions continue to determine what the underlying layer can access. The decision that remains is which clients your security team approves to reach that endpoint — worth scoping early, since it typically involves the same review process as any other data integration.

Genie One is the business experience, not an analytics back end

It is easy to assume a data assistant only handles questions about numbers. The native experience covers considerably more of the working day — which is why it is the recommended starting point rather than something sitting behind another tool.

💬
Ask a business questionIn plain language, grounded in your governed definitions.
📊
Explore a governed dashboardOpen and interrogate the certified assets your teams already built.
📝
Draft a business documentTurn a grounded answer into shareable written work.
🔄
Schedule a recurring taskHave the answer arrive on a cadence instead of being re-asked each time.
🎯
Use a scoped Genie AgentWork with a curated agent tuned to one domain and its rules.

Confirm feature availability, connector configuration, permissions, and approval requirements for your own deployment.

Optional extension Extend Genie when another approved tool needs it

If a team is already standardized on an external assistant, connect that approved client through the Genie One MCP server. The request then draws on the same governed context and answering capability, inside the same permissions.

Genie OntologyUnity Catalog governed
→
Genie One MCP
→
ClaudeChatGPTCursorYour own app
What this is, precisely. An MCP connection exposes governed answering capability. It is not a bulk export of your context, and it is not a permission bypass. External assistants have their own persistent context and reusable artifacts — the argument is not that they forget. It is that on their own they do not establish your Databricks-governed semantics, source authority, or shared ontology. The context investment stays centered in Databricks; the external tool becomes one more place it can be consumed.
5

One governed layer, every assistant

The assistant your teams prefer is not the thing that determines whether an answer is right. Your context layer is. Here is what that looks like in practice — and why what you build becomes an asset you own.

The same question, asked twice

One model, one prompt, one person asking. The only variable is whether your governed layer is reachable.

Claude no governed layer
JD
What is our same-store NOI this quarter?
✦
I don’t have access to your revenue data. If you paste the figures I can calculate it — though I’ll also need to know how your organization defines same-store NOI: which assets qualify as same-store, whether management fees sit in operating expenses, and how concessions and free rent are treated.
Awaiting input Correct only if the person pasting remembers every rule
Claude Genie MCP connected
JD
What is our same-store NOI this quarter?
⚙Calling genie_one · resolving against your governed definitions
✦
+3.4%
Same-store NOI for Q1, using the approved definition and excluding assets acquired mid-period or under renovation.
via Genie MCP Your permissions applied Source: your certified NOI asset
The assistant on its own
The same assistant, reaching your layer
Nothing about the model changed. It reasons identically on both sides. What changed is that one of them could reach definitions your business had already agreed on — which is why the layer, not the window, is the part worth investing in.

The same answer, and everything you do with it

Ask it natively and you get the identical figure — deliberately identical, because it came from your layer rather than the surface. What the native experience adds is the work that follows the number.

Genie One your workspace
JD
What is our same-store NOI this quarter?
⚡
+3.4%
Same-store NOI for Q1, using the approved definition and excluding assets acquired mid-period or under renovation.
Genie Ontology Your permissions applied Source: your certified NOI asset
And from the same place, without switching tools
📊
Open the governed dashboard behind itInterrogate the certified asset the number came from.
📝
Draft the board paragraphTurn the grounded answer into shareable written work.
🔄
Schedule it monthlyHave it arrive on a cadence instead of being re-asked.
🎯
Hand it to a scoped Genie AgentLet a curated agent own this domain and its rules.
🔍
Inspect the citationsSee exactly which sources the answer drew on.

The answer is not the asset

A number answers today’s question. What makes the next one cheaper is the definition underneath it — and that is the thing you are actually building.

Same-store NOI governed definition
ConceptSame-store NOI against the prior comparable period
CalculationYour approved formula, agreed by the owning team
Plan and periodYour fiscal calendar and the approved plan version
SourceYour authoritative governed asset, not whichever file was nearest
AccessThe requesting person’s own permissions, applied every time
EvidenceCitations to the sources actually used in the answer
The reusable asset is the business meaning — not just the last answer.

What one governed question actually buys you

One question followed end to end — and the two things left over once it has been answered.

❓ “Which assets are tracking below underwritten NOI, and what explains the variance?”
01
Asked in ClaudeThe assistant recognizes the question needs enterprise data it does not hold.
02
Routed through the Genie MCP serverNo new connector to build, and no second permission model to keep in step with the first.
03
Genie Ontology resolves the meaningWhich NOI definition is approved — cash or GAAP, before or after capital reserves — which underwriting version applies, whether management fees sit in operating expenses, and how assets roll up to fund and joint venture.
04
OntoRank selects the authorityA certified metric view outranks a stale notebook, because authority weighs source, usage, and freshness — not keyword relevance alone.
05
Unity Catalog gates the readThe query runs inside the asker’s own permissions, so fund and joint-venture segregation holds automatically. An analyst supporting one fund sees that fund; a JV partner sees only their own deals.
06
The answer returns with its proofThe figure, the SQL behind it, and citations to the sources actually used — so it can be checked rather than trusted blindly.
What you now own
The three assets set out in Section 1 — a certified definition, a trusted query, and a curated agent — now exist because one person asked one question in Claude. Nobody set out to build them.
Built once, then inherited by whoever asks next.
What you did not have to duplicate
Connector configuration Permission design Metric definitions Prompt instructions Monitoring and review Source reconciliation
Without a shared layer, every AI tool you add needs its own copy of all six. Each copy is another place the definitions can quietly drift apart — and another line of spend doing work you have already paid for once.

You own this layer. That is the part that lasts.

Every definition your teams certify lives in your Unity Catalog, governed by your permissions, cited in your answers. It is not held inside a vendor’s conversation history, and it is not reassembled from scratch in each person’s prompt.

That is what makes it an asset rather than a subscription feature. Models will keep changing — faster, cheaper, better — and your teams may well prefer a different assistant a year from now. The layer stays yours either way. Swapping the window does not cost you the business meaning underneath it.

Which is the real argument for building here: not that one chat interface beats another, but that the understanding of your business becomes something you accumulate and keep, instead of something every employee reconstructs on every prompt.

6

What it costs, and what you stop paying twice for

Two halves of one question: the costs that fall as the layer matures, and the spend you can finally see in one place.

📉

1. Right-sizing AI capacity

Where premium general-purpose AI tiers were purchased because data questions exhausted message or token limits, moving that work to a purpose-built governed surface can reduce the capacity required for those users.

♻

2. Declining cost per answer

Each reviewed dashboard, curated Agent, and approved definition reduces future rework. As reuse replaces repeated ad hoc analysis, the effort and cost of answering a recurring question falls.

🔧

3. Consolidated integration effort

One governed MCP layer is less work to build, secure, and maintain than a separate data connector, permission design, and review cycle inside each AI tool you adopt.

84.5%
First-attempt accuracy
in roughly
8 minutes

Accuracy and latency are cost questions. A wrong answer gets rebuilt by hand, and a slow one blocks whoever is waiting on it. On a 28-question internal enterprise-data-analysis suite run in June 2026, Genie answered 84.5% correctly on the first attempt, in roughly eight minutes.

The strongest general-purpose coding agent evaluated alongside it — given the same Databricks SQL MCP access — reached 52.4%. The others evaluated came in at 40.5% and 25.0%, which is what makes 52.4% the strongest rather than a convenient midpoint. Runtime told a similar story — roughly 8 minutes for Genie against 15 and 24 minutes for two of the agents, though one was faster at 7 minutes while being far less accurate. Those agents run frontier models, so the gap is not reasoning quality. It is the context layer underneath, which is also why a result like this is reproducible in your environment rather than dependent on how well each person prompts.

A limited Databricks vendor benchmark on a 28-question internal suite, June 2026. Competing agents are anonymized; this is not a comparison against any named assistant, and not a customer guarantee. Your own results will depend on how your definitions and assets are curated. Published disclosure: Introducing Genie One, Genie Ontology, and Genie Agents.

How the commercial structure supports this

  • Connecting an AI tool you already own costs nothing extra. There is no separate charge for the Genie One MCP beyond the underlying Genie usage, so a team working through Claude or Copilot against your governed layer is not paying twice for the same answer.
  • Consumption is pay-as-you-go, with limits you set. Administrators can configure alerts and hard caps per user and per workspace, so a pilot cannot quietly overrun its budget.
  • Spend is recorded in Unity Catalog system tables rather than a vendor portal — which is what makes the attribution described in the next section possible.
  • Building the learned ontology is not currently metered. Treat this as current rather than permanent, and confirm the terms that apply to you with your Databricks account team.

Comparing the two operating models

A qualitative view of where effort and spend accumulate.

Category Without a governed Genie layer With Genie One, Genie Code, and MCP
Business-user data access General-purpose AI seats used as an informal analytics tool, often at premium tiers Genie One provides a purpose-built front door for governed data questions
Analyst effort Repeated context assembly, schema explanation, and re-answering questions already answered Genie Code converts recurring questions into shared assets that Genie One surfaces first
Integration maintenance A separate connector and review cycle inside each AI tool adopted One governed layer exposed once through MCP
Governance overhead Semantic logic distributed across prompts, spreadsheets, and individual knowledge Definitions, permissions, and lineage centralized in Unity Catalog
Overall trajectory Variable AI cost and inconsistency both tend to grow with adoption Incremental cost for data work declines as the governed asset base matures

What to measure during a pilot

  • Time from question asked to reviewed, trusted answer
  • Proportion of questions answered by an existing governed asset rather than new analysis
  • Number of recurring questions converted into reusable assets
  • AI usage and compute consumption attributable to governed data work
  • Instances where two tools previously produced different figures for the same metric

Cost outcomes depend on your existing licensing, data readiness, and usage patterns, so these levers are best quantified with your own figures rather than generic benchmarks.

Eliminating redundant spend — and seeing all of it in one place

AI cost tends to fragment before anyone notices. Consolidation and visibility address different halves of that problem.

Where redundant AI spend accumulates

Paying for the same model more than once

A team subscription for an assistant, the same model family called directly through an API in a custom application, and the same provider served again through the data platform. Three commercial relationships, one underlying model.

Overlapping tool footprints

Two or three assistants each partially covering data questions, with none of them authoritative. Teams pay for all of them and still escalate to an analyst for the number that has to be right.

Decentralized purchases with no attribution

Individual and departmental AI subscriptions that never appear in a consolidated view until renewal. Without central reporting, duplicate coverage is invisible.

Duplicated integration and monitoring

Separate connectors, logging, spend tracking, and security review inside each tool. The same governance work is performed repeatedly, once per vendor.

Repeatedly answering the same question

Without a governed asset to reuse, each occurrence of a recurring question consumes analyst time and AI capacity again from the beginning.

Capacity provisioned above actual need

Premium tiers purchased organization-wide because a subset of users needed the headroom, leaving occasional users on capacity they rarely consume.

Unity Gateway: one control plane for AI cost and usage

Because every request routes through the gateway, cost and usage questions have a single answer rather than one answer per tool.

Unity Gateway — shown as AI Gateway in workspace navigation — is the control plane for AI traffic, not only its meter. It governs which models and MCP services an agent may reach, applies guardrails, approval requirements, and rate limits, and records every request against Unity Catalog system tables. That answers the questions scattered tooling cannot: how many agents are running and who owns them, why a request was blocked, whether sensitive data reached a third-party model, and what each team actually spent.

AI traffic across your organization
⚡Genie
🤖External models
🔧Coding agents
⚙️Custom agents
📜ai_query in SQL
📦Applications
↓

Unity Gateway

Agent Registry · Access Control · Contextual Policies · Budgets · Smart Routing · Agent Monitoring

Cost by service and model Cost by user or service principal Cost by team tag Token and latency detail
↓
Unity Catalog system tables — the shared record behind the built-in usage dashboard and Governance Hub

The system tables behind the numbers

system.ai_gateway.usage Per-request usage metrics including token counts, latency, and requester details for everything routed through the gateway. Also captures ai_query calls to Databricks-provided model services.
system.billing.usage Billable consumption in DBUs. Gateway traffic is enriched with the model service name, destination model, requesting principal, and any service tags — which is what makes attribution possible.
system.billing.list_prices List pricing, joined to consumption records to estimate spend on Databricks-hosted foundation models.
system.ai_gateway.external_model_spend Estimated spend for requests routed to external model providers, aggregated hourly.

Questions you can answer from a single place

Each of these typically requires a separate export or vendor portal without a gateway in front of your AI traffic.

Question Where the answer comes from Why it matters
Which models are we actually using, and at what volume? Usage dashboard, Requests and Token Usage views Reveals overlapping coverage and models retained but no longer used
What does each team or project spend on AI? Service tags on billing records, grouped by team or cost centre Moves AI from an unallocated central cost to an attributable one
Which users or service principals drive the most consumption? Requesting principal on usage and billing records Distinguishes genuine heavy use from a misconfigured job or runaway agent
Is spend trending in a way we should act on? Cost Over Time and Cost Breakdown views Surfaces trends early rather than at invoice or renewal
Are we paying an external provider for something available internally? External model spend alongside Databricks-hosted consumption Directly identifies duplicated model coverage
What exactly was sent to a model, and by whom? Request and response logging to Unity Catalog tables Supports audit, debugging, and review without a vendor-side export
How do we prevent a single workload consuming a quarter of budget? Rate limits applied per model service A preventive control rather than a retrospective finding

Consolidation and visibility solve different problems

Routing AI traffic through one governed layer reduces how many overlapping tools and integrations you pay to maintain. Gateway observability tells you which of the remaining ones earn their cost. Both are needed: consolidation without measurement is a guess, and measurement without consolidation only documents fragmentation more precisely.

What to confirm for your environment

  • Unity Gateway requires a supported region, and the billable usage system table must be enabled for your account.
  • System table access requires account and metastore admin roles by default. Admins can publish the usage dashboard to share analysis without granting direct table access.
  • External model spend is estimated from each provider's published prices and is informational rather than a billing record.
  • Token and cost metrics are not available for every provider or model, so some charts may be empty depending on your configuration.
  • Cost attribution by team or project depends on service and request tags being applied consistently — worth agreeing on a tagging convention before rollout.
7

Getting started: turn it on, then prove it

Two phases, in order. Enabling Genie One is a configuration task; proving it is a short, measured pilot against a baseline you record first.

Phase 1Turn it on

No new platform, no migration, no separate licence

Genie One is the simplified Databricks interface for business users. It runs on the workspace and Unity Catalog permissions you already have — so the work is granting access and publishing the right assets behind it.

Setup in five steps

1
Workspace admin

Grant the Consumer access entitlement

Consumer access is the least-privileged workspace entitlement and is sufficient for Genie One chat. Higher entitlements such as Databricks SQL access, Workspace access, or workspace admin also qualify, so existing users may already be covered. Assign it to groups rather than individuals so access follows your existing identity model.

2
Workspace admin

Start at the workspace level

Workspace-level Genie One is the path for chat today: it is scoped to a single workspace and its members, and users see only the assets explicitly shared with them or with a group they belong to. Assign the entitlement per workspace for the teams taking part in your pilot.

3
Data team

Publish the assets users should land on

Genie One surfaces the AI/BI dashboards, Genie Agents, and Databricks Apps that have been shared with a user or with a group they belong to. Share at least one reviewed dashboard before opening access — an empty home page is the most common reason early adoption stalls.

4
Data team

Confirm how queries run against your permissions

Every Genie One interaction respects Unity Catalog permissions. When sharing an asset, the option to run queries using the viewer's credentials means each person sees only the data their own Unity Catalog privileges allow. Where external knowledge sources such as SharePoint or Google Drive are connected, those sources' permissions are honoured as well.

5
Everyone

Send users the link

Users whose only entitlement is Consumer access land on Genie One automatically after signing in. Anyone with additional entitlements can switch using the app switcher in the upper-right corner of the workspace, or navigate directly.

Direct URL
https://<your-workspace-url>/one
Append /one to your workspace URL. Users with only Consumer access are directed here on sign-in, so this link is mainly for colleagues who also have workspace or SQL entitlements.

Optional extensions worth enabling

📱

Mobile access

The Genie mobile app supports iOS and Android and is enabled by default for accounts in Public Preview. If it is not already on, an account admin can enable Genie One Mobile from Previews in the account console.

📊

Microsoft Excel

With the Databricks Excel Add-in configured, users can open Genie One from the add-in pane and ask questions in plain language without leaving the workbook — answers include citations and conversation history.

✨

Personalized starter questions

When enabled, the home page offers one-click prompts tailored to recent activity, which lowers the barrier for first-time users. Currently in Beta and controlled by workspace admins from the Previews page.

Documentation

Links point to Azure Databricks documentation. Equivalent pages for AWS and GCP deployments are available at docs.databricks.com. Features marked Beta or Public Preview may vary by region and account configuration.

A note on naming

Genie One was previously called Databricks One. Older internal documentation, training material, or release notes may still use the former name.

Phase 2Prove it

Running your proof of concept

Start with one team and a few recurring workflows, measured against a baseline you recorded first.

1
Sponsor + team lead

Pick one team and a few recurring workflows

That might be Legal reviewing incoming contracts, Finance producing a monthly budget, or Sales preparing account briefs. Choose work the team already does often, where a better answer or a faster process would make a real difference.

2
Team lead + data owner

Define your benchmarking questions

Representative questions with known-good answers — including the definitions, sources, and reasoning Genie should use. They give you a concrete way to evaluate accuracy before rollout and to track whether Genie is improving over time. Include straightforward questions, deliberately ambiguous ones, and cases where permissions should produce different answers for different users.

3
Team lead

Record today’s baseline, before Genie

How accurate is the existing process? How long does it take? Where do people get stuck? Capturing this first is what later lets you show improvement rather than assert it — and it is the step most teams skip.

4
Domain experts

Bring in the people who own the definitions

For a Finance workflow this is typically the FP&A lead who owns monthly reporting and knows exactly how the company defines budget and actuals. They help codify the metrics, terms, and concepts Genie needs to understand, and certify the authoritative context through metric views, pages, and domains. They are often keen to see the user feedback as it starts to arrive.

5
Data team

Certify what you already have

This is not a cold start. You almost certainly hold useful knowledge already — in high-quality Genie Agents, dashboards, queries, and notebooks. Certify those assets and enable ontology snippets so Genie can learn from that existing work, and make sure the users in your pilot have access to them so the relevant snippets can actually surface.

6
Data team + IT

Connect the sources the workflow needs

Does it need SharePoint? Gmail? Jira? Slack? Make sure people can read from and write to the third-party tools the work depends on. You can add instructions where they help, but you probably will not need many to begin with — if the underlying semantics and source material are strong, Genie can do a great deal without a long list of rules.

7
Everyone

Put Genie in the team’s hands, then iterate

See where it succeeds, where it gets confused, and where the ontology needs more context. Refine the semantics and instructions as you go — and once it is working well, expand to the next workflow, the next domain, and the next team.

Why benchmarking questions matter more than a launch date

A pilot without known-good answers can only tell you whether people liked it. With benchmarking questions and a recorded pre-Genie baseline, you can state plainly how accuracy and cycle time changed, and show the ontology improving over successive passes. That evidence is what turns one team’s pilot into an organization-wide rollout.