{"slug": "put-copilot-metadata-where-your-team-can-review-it", "title": "Put Copilot Metadata Where Your Team Can Review It", "summary": "As Microsoft Copilot in Power BI uses semantic model metadata to answer business questions, this metadata must be treated as critical engineering assets—reviewed, versioned, and stored in source control alongside model changes. It details how Copilot metadata (including instructions, verified answers, and schema) is stored in a dedicated `Copilot/` folder within PBIP projects, and warns that migrating from older Q&A tooling permanently disables Q&A features, requiring careful planning. The key takeaway is that teams should review Copilot metadata changes in pull requests to ensure AI context aligns with the business and the semantic model.", "body_md": "Originally published at https://shai-kr.github.io/data-ninja-ai-lab/blog/2026-05-20-copilot-metadata-source-control.html\nCopilot in Power BI is moving the semantic model into a new role.\nIt is no longer only the layer that stores measures, relationships, tables, and security rules.\nIt is also becoming the place where teams define how AI should understand the business.\nThat changes the engineering standard.\nIf Copilot uses metadata to answer business questions, that metadata should not live as invisible configuration edited by one report author on a good day.\nIt should be reviewed.\nIt should be versioned.\nIt should be tested.\nAnd in serious Power BI environments, it belongs in source control.\nThe easy version of this story is:\nPower BI now has tools to prepare semantic models for AI.\nThat is true, but it is not the useful lesson.\nThe more important shift is that Copilot readiness is becoming part of the semantic model lifecycle.\nPower BI now gives model owners tools such as:\nThis is not report polish.\nThis is semantic engineering.\nThe detail that matters most for engineering teams is simple:\nWhen a Power BI project is saved as PBIP, Copilot metadata is stored in a Copilot/\nfolder inside the semantic model project.\nThe structure looks roughly like this:\nPBIP/\n├── Model.SemanticModel/\n│ ├── definition/\n│ ├── Copilot/\n│ │ ├── Instructions/\n│ │ │ ├── instructions.md\n│ │ │ ├── version.json\n│ │ ├── VerifiedAnswers/\n│ │ │ ├── definitions/\n│ │ │ ├── version.json\n│ │ ├── schema.json\n│ │ ├── examplePrompts.json\n│ │ ├── settings.json\n│ │ └── version.json\n│ └── definition.pbism\nThat means Copilot behavior can become part of the same workflow as model changes.\nA pull request can now include:\nThose changes should be reviewed together because users experience them together.\nMost Power BI teams already know that bad semantic models create bad reports.\nCopilot adds another layer:\nBad semantic metadata creates bad AI answers.\nA vague measure name is no longer just annoying for developers.\nA weak description is no longer just documentation debt.\nA poorly governed synonym or verified answer can steer users toward the wrong interpretation of the business.\nThis is especially important for models used across multiple reports or apps. When the model becomes approved for Copilot, the risk is not isolated to one page. It can affect every experience that depends on that semantic model.\nIf a team is using PBIP and Git, Copilot metadata should not be treated as noise in the repo.\nAt minimum, I would review these changes before promotion:\nDo the instructions reflect how the business actually talks about the data?\nIf finance calls it “net revenue” and sales calls it “bookings”, the model needs to make that distinction clear.\nA verified answer should be tied to a known, trusted interpretation.\nIt should not become a workaround for a weak model.\nThe phrasing should match real user language, not developer language.\n“Show customer churn by cohort” and “why did retention drop last month?” are not the same question.\nCopilot schema updates should be reviewed with table, column, relationship, and measure changes.\nOtherwise the AI context can drift away from the model.\nSome tooling changes require a model refresh after deployment before they sync into the service.\nThat should be part of the deployment checklist, not something discovered after users start asking questions.\nThere is another detail teams should not ignore.\nPower BI can migrate older Q&A tooling metadata into the newer Copilot tooling format, but this is not a harmless cosmetic migration.\nAfter the upgrade, Q&A features are permanently disabled for the model and connected reports. Some Q&A metadata migrates, such as synonyms and suggested questions. Other metadata does not.\nThat means the migration should be planned like a model change.\nNot like a dialog box someone clicks through.\nBefore upgrading a production model, I would check:\nThe best pattern is not complicated.\nFor production semantic models, treat AI readiness as a normal engineering workflow:\nCopilot/\nfolder in GitThis gives the team a simple standard:\nIf the change can affect what Copilot tells a user, it deserves the same discipline as a DAX or model change.\nFor a Power BI team using Copilot, I would add these checks to the semantic model review process:\nCopilot/\nchange?The real shift is not that Power BI added more AI buttons.\nThe real shift is that AI behavior is becoming part of the analytics system.\nThat means BI teams need to stop treating Copilot readiness as a last-mile report setting.\nIt belongs with the semantic model.\nIt belongs in Git.\nIt belongs in code review.\nAnd it belongs in the same quality process as every other change that can affect business interpretation.\nThat is how Power BI teams move from “we turned Copilot on” to “we can trust what Copilot is allowed to say.”", "url": "https://wpnews.pro/news/put-copilot-metadata-where-your-team-can-review-it", "canonical_source": "https://dev.to/shai_karmani_2521c2f8e837/put-copilot-metadata-where-your-team-can-review-it-5f0o", "published_at": "2026-05-20 23:46:48+00:00", "updated_at": "2026-05-21 00:02:43.207875+00:00", "lang": "en", "topics": ["data", "artificial-intelligence", "enterprise-software", "developer-tools"], "entities": ["Power BI", "Copilot", "Microsoft"], "alternates": {"html": "https://wpnews.pro/news/put-copilot-metadata-where-your-team-can-review-it", "markdown": "https://wpnews.pro/news/put-copilot-metadata-where-your-team-can-review-it.md", "text": "https://wpnews.pro/news/put-copilot-metadata-where-your-team-can-review-it.txt", "jsonld": "https://wpnews.pro/news/put-copilot-metadata-where-your-team-can-review-it.jsonld"}}