# Microsoft IQ Series: Video Episodes and Jupyter Cookbooks for Foundry IQ, Work IQ and Fabric IQ > The IQ Series is Microsoft's hands-on learning repository for its unified intelligence layer, pairing weekly video episodes with Jupyter notebook cookbooks and Azure deployment templates. It is a teaching resource, not a library you import. **microsoft/iq-series** — The IQ Series is a hands-on learning experience for Microsoft IQ: Microsoft's unified intelligence layer for the enterprise, spanning Foundry IQ, Work IQ, and Fabric IQ. The series includes video episodes, Jupyter notebooks, and Azure deployment templates. - Repository: https://github.com/microsoft/iq-series - Website: https://aka.ms/iq-series - Stars: 380 · Forks: 267 - Language: Jupyter Notebook - License: MIT - Published: 2026-09-10 · Updated: 2026-09-10 · Language: en - Canonical page: https://hysenlabs.com/projects/microsoft-iq-series ## What the IQ Series actually is, and who it is written for The repository is a curriculum, not a package. Microsoft IQ is described in the README as "Microsoft's unified intelligence layer for the enterprise," composed of three services: Foundry IQ, a managed knowledge layer that connects structured and unstructured data across Azure, SharePoint, OneLake and the web so agents can reach permission-aware knowledge; Work IQ, which exposes organizational context, relationships and work patterns to agents; and Fabric IQ, which unifies business semantics across data, models and systems. The IQ Series teaches those three through video episodes and companion notebooks. The audience is narrow and specific. You are expected to be comfortable in Azure, to run Jupyter notebooks, and to care about retrieval quality rather than about model training. The repository's topics list confirms the framing: rag, knowledge-base, knowledge-source, microsoft-foundry, microsoft-fabric, copilot-api, github-copilot. If you are evaluating a vector database or building a document pipeline from scratch, this is not aimed at you. If you are trying to understand where Microsoft's enterprise retrieval layer sits relative to plain RAG, that is exactly the question the first Foundry IQ episode poses in its title, "Unlocking Knowledge for your Agents." One structural detail matters more than it looks. The README says the series "kicks off with Foundry IQ episodes, followed by Work IQ episodes," with Fabric IQ content coming soon. So the repository is asymmetric: two thirds of the advertised subject matter is documented, one third is a placeholder row in the episode table. ## How the episodes, cookbooks and infra folder fit together The layout is shallow and predictable, which is a genuine virtue in a teaching repo. Three top-level content directories (Foundry-IQ/, Work-IQ/, Fabric-IQ/) sit alongside images/, infra/, .devcontainer/, .github/ and .vscode/. Each episode lives in its own numbered folder with a README, and each of those folders contains a cookbook/ subdirectory. Foundry IQ has three published episodes: core components and agent architecture, building the data pipeline with knowledge sources, and querying multi-source knowledge bases. Work IQ has three as well, covering architecture and protocol strategy across REST, A2A and MCP; the A2A protocol and agent discovery patterns; and tooling with a unified MCP server used from Work IQ CLI and GitHub Copilot CLI. That progression is the actual mechanism of the series. Episode one explains a component. Episode two moves data into it. Episode three queries it. The Work IQ track follows the same shape but swaps the subject from knowledge retrieval to protocol and tooling, which is why the episode descriptions mention REST, A2A and MCP in the same sentence. If you read the two tracks side by side, the split is clear: Foundry IQ is about what the agent knows, Work IQ is about how the agent reaches the organization. The infra/ directory is the part most readers will underestimate. The repository description states the series includes "Azure deployment templates," and infra/ is where those live. For a reader who wants to run a cookbook against real resources rather than read it, infra/ is the entry point, and the README does not walk through it. That is a documentation gap worth knowing about before you start. The episode format is fixed: roughly one minute of introduction, about fifteen minutes of technical discussion between the product group and advocacy, and a one-minute close-out. That cadence tells you what the videos are for. They are orientation, and the notebooks carry the detail. ## Running your first Foundry IQ cookbook The README does not publish install commands. It points to the episode folders, each of which holds a cookbook or markdown lab instructions, and it lists a .devcontainer/ directory at the repository root. The practical path is to clone the repository, open the episode folder you want, and let the dev container definition supply the environment, or reproduce it yourself in a Jupyter environment. Start by getting the code and locating the first cookbook: ```bash git clone https://github.com/microsoft/iq-series.git cd iq-series/Foundry-IQ/1-Foundry-IQ-Unlocking-Knowledge-for-Agents/cookbook ``` From there, the notebook itself is the instruction set. The README does not document which environment variables or Azure resources each notebook expects, so the first cell of the notebook is the authoritative source. Open it and read the setup cell before running anything. If you prefer to work from the repository's own container definition rather than your local Python, the .devcontainer/ entry at the root is the supported route: ```bash cd iq-series code . ``` Visual Studio Code will offer to reopen the folder in the container when a .devcontainer/ configuration is present. The README does not state which base image or extensions that configuration uses, so treat the container as a convenience, not as a documented guarantee. The badge process doubles as a completion check. The README asks you to fork the repository, save your notebook outputs in your fork, capture a final output screenshot for each episode, submit a badge request issue, and complete the badge form. Badges are issued by the Global AI Community, so an account there is required. If you finish all three Foundry IQ cookbooks, that sequence is the confirmation that your environment produced real output rather than a skipped cell. ## Where the IQ Series stops being the right tool The most obvious limitation is stated in the README itself: Fabric IQ content is coming soon. Anyone whose question is about business semantics across data, models and systems will find an empty row in the episode table. That is a third of the advertised scope. A second limitation is the nature of the artifact. There are no releases in the repository, and the top-level entries are documentation, notebooks, images and templates. Nothing here is a versioned library you add to a dependency file. If your team needs a pinned, supported SDK with a semantic version and a changelog, the IQ Series cannot supply it, because it does not publish one. You would be reading notebooks to learn an interface whose stability the repository never claims. A third is the environment assumption. The cookbooks assume Azure resources and, for the Work IQ track, Microsoft 365 context. Work IQ is described as understanding "context, relationships, and work patterns," which is organizational data by definition. A solo developer with a personal tenant and no SharePoint or OneLake content will get less out of episodes two and three of the Foundry IQ track than someone inside a company with real corpora. The series rewards readers who have data to point it at. Finally, the video schedule is a dependency. Foundry IQ episodes premiered weekly on Wednesdays at 9 AM PT starting March 18, 2026, and Work IQ episodes premiered at 9 AM PT on June 2, 2026. The repository is a companion to a broadcast cadence, so its completeness tracks the release schedule rather than a documentation plan. ## How it differs from Microsoft's own sample repositories and from plain RAG stacks The nearest comparison is not a competitor product but a different kind of resource: the Microsoft IQ Deep Dive, which the README points to as a complementary learning experience. The README describes it as "a three-day workshop concept with Python notebooks and agents." The difference in approach is format and depth. The IQ Series is episodic and incremental, roughly seventeen minutes per episode with a cookbook attached, designed to be consumed one Wednesday at a time. The Deep Dive is a concentrated multi-day workshop. If you need to brief a team next week, the workshop structure fits better. If you are learning alongside a weekly release and want each concept demonstrated before the next one lands, the series fits better. They are not substitutes; the README presents them as complements. The other comparison is architectural, and the README makes it explicitly. Microsoft IQ is positioned as going "beyond traditional RAG for true enterprise intelligence." The concrete difference named in the description is permission-aware knowledge: Foundry IQ connects data across Azure, SharePoint, OneLake and the web so agents can access knowledge that respects existing permissions. A conventional RAG stack built on a vector store has no native notion of the SharePoint permission model, and you build that enforcement yourself. That is the actual distinction, and it is the reason the series spends an entire episode on knowledge sources and another on multi-source query paths rather than jumping straight to prompts. Work IQ's differentiator is protocol breadth. The episode list names REST, A2A and MCP, and episode three covers a unified MCP server used from Work IQ CLI and GitHub Copilot CLI. If your integration surface is MCP, that episode is the relevant one; if it is REST, episode one covers that ground. ## Licence, maintenance and what an upgrade costs you The repository is MIT licensed, per its LICENSE file and the repository metadata. MIT is permissive: it allows reuse, modification and redistribution with the licence and copyright notice retained. That matters here because the notebooks are meant to be forked, and the badge instructions explicitly ask you to fork the repository and save your outputs in your fork. Nothing in the repository states a separate licence for the video content or the badge artwork, so if you plan to reuse images or episode material beyond the code, check the source rather than assuming MIT covers it. This is not legal advice; read LICENSE and TRADEMARKS.md, both of which are present at the repository root. The last push to the default branch was on 2026-08-03. The repository is not archived. Because the series is tied to a weekly episode cadence and a "coming soon" Fabric IQ track, expect the content directories to keep moving. The upgrade cost is therefore not a dependency bump; it is re-reading notebooks. If you forked the repository to earn a badge, your fork diverges from upstream as episodes are added, and merging later changes means resolving notebook conflicts, which are unpleasant in JSON-based .ipynb files. There is no release history to consult, so you cannot tell from tags which episode content changed and when. The only signal is the commit history on main. For a teaching repository that is acceptable. For anything you intend to build on, it is the reason to treat the notebooks as reference implementations rather than as a base to fork and maintain long term. ## Conclusion Adopt the IQ Series if you are an Azure or Microsoft 365 engineer who needs working notebooks before committing to Foundry IQ or Work IQ, and if you can absorb the cost of a fast-moving preview surface. Skip it if you need a supported SDK, a stable API contract, or Fabric IQ material today, because the README lists Fabric IQ as coming soon and no release artifacts are published. Before you start, open the cookbook folder for the episode you care about and confirm which Azure resources and credentials its first cell expects, then check the .devcontainer/ configuration to see whether the pinned environment matches your own. ## FAQ ### What is Microsoft IQ in the IQ Series? The README describes Microsoft IQ as Microsoft's unified intelligence layer for the enterprise, bringing together three services: Foundry IQ, Work IQ and Fabric IQ. Together they are positioned as letting agents reason, retrieve and act with business context beyond traditional RAG. ### Which parts of the Microsoft IQ Series are published so far? Foundry IQ and Work IQ each have three published episodes with cookbooks, while the episode table lists Fabric IQ as coming soon. The README notes the series kicks off with Foundry IQ episodes followed by Work IQ episodes. ### How do I earn the Foundry IQ badge from the Microsoft IQ Series? Complete all three Foundry IQ cookbooks, fork the repository and save your notebook outputs in the fork, capture a final output screenshot per episode, submit a badge request issue, and complete the badge form. Badges are issued by the Global AI Community, so you need an account there first. ### Is the Microsoft IQ Series a library I can install? No. The repository holds video episode links, Jupyter notebook cookbooks, markdown lab instructions, images and Azure deployment templates under infra/, and it publishes no releases. The README does not give an install command, so you clone the repository and run the notebooks. ## Sources - [Issues](https://github.com/microsoft/iq-series/issues) - [License: MIT](https://github.com/microsoft/iq-series/blob/main/LICENSE) - [microsoft/iq-series on GitHub](https://github.com/microsoft/iq-series) - [Project website](https://aka.ms/iq-series) - [README](https://github.com/microsoft/iq-series/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/microsoft-iq-series