<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>GiGi Levens</title>
    <link>https://gigilevens.com/</link>
    <description>Field notes on data, AI, and the craft of governance — Microsoft Fabric, Power BI, Purview, Copilot.</description>
    <language>en</language>
    <lastBuildDate>Mon, 07 Sep 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://gigilevens.com/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>The Belated Recap: The Data &amp; AI Releases I Missed on Sabbatical</title>
      <link>https://gigilevens.com/writing/belated-recap-releases-i-missed-on-sabbatical/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/belated-recap-releases-i-missed-on-sabbatical/</guid>
      <pubDate>Mon, 07 Sep 2026 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>While I was away, Microsoft held two of the year&#x27;s biggest data events — FabCon in March, Build in May — and reshaped the roadmap around agents. I read six months of announcements in a weekend so you don&#x27;t have to. Here&#x27;s what actually matters, assessed late and on purpose.</description>
    </item>
    <item>
      <title>Fabric in Production: Governance Patterns That Actually Hold</title>
      <link>https://gigilevens.com/writing/fabric-in-production-governance-patterns-that-actually-hold/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-in-production-governance-patterns-that-actually-hold/</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Everyone can stand up a Fabric workspace. Far fewer can run fifty of them without the whole estate turning into a swamp. The topology, capacity, and access patterns I&#x27;ve watched survive contact with production — and the ones that quietly fall apart.</description>
    </item>
    <item>
      <title>The Governance Debt Behind Every Stalled AI Project</title>
      <link>https://gigilevens.com/writing/the-governance-debt-behind-every-stalled-ai-project/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/the-governance-debt-behind-every-stalled-ai-project/</guid>
      <pubDate>Mon, 10 Aug 2026 00:00:00 +0000</pubDate>
      <category>Data Governance &amp; Compliance</category>
      <description>The AI pilot that never reaches production usually didn&#x27;t fail on the model. It failed on ten years of data-governance shortcuts that all came due at once. A field guide to the debt, and how to pay it down without a two-year programme.</description>
    </item>
    <item>
      <title>Back From a Long Pause: What I&#x27;m Watching in Data &amp; AI for H2</title>
      <link>https://gigilevens.com/writing/back-from-sabbatical-what-im-watching-h2/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/back-from-sabbatical-what-im-watching-h2/</guid>
      <pubDate>Mon, 13 Jul 2026 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>The break ran longer than the itinerary said it would. Now that I&#x27;m easing back to the desk, here&#x27;s the short list of what I&#x27;m actually watching in data and AI for the second half of the year — and what I&#x27;ve stopped watching.</description>
    </item>
    <item>
      <title>Stepping Away for a While</title>
      <link>https://gigilevens.com/writing/stepping-away-for-a-while/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/stepping-away-for-a-while/</guid>
      <pubDate>Sun, 01 Mar 2026 00:00:00 +0000</pubDate>
      <category>Craft &amp; Communication</category>
      <description>A short note before I go quiet: I&#x27;m taking an extended break to travel with my family. The blog pauses until the summer — on purpose.</description>
    </item>
    <item>
      <title>AI Readiness Is Just Data Readiness With Better Marketing</title>
      <link>https://gigilevens.com/writing/ai-readiness-is-just-data-readiness-with-better-marketing/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/ai-readiness-is-just-data-readiness-with-better-marketing/</guid>
      <pubDate>Mon, 16 Feb 2026 00:00:00 +0000</pubDate>
      <category>AI &amp; Responsible Tech</category>
      <description>Every organisation wants to be &#x27;AI-ready&#x27; this year, and a whole industry has appeared to sell it to them. Strip the branding and the checklist is almost identical to the data-readiness one we&#x27;ve been ignoring for a decade. That&#x27;s either deflating or liberating, depending on how you take it.</description>
    </item>
    <item>
      <title>What 13 Years in Data Taught Me About Hype Cycles</title>
      <link>https://gigilevens.com/writing/what-13-years-in-data-taught-me-about-hype-cycles/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/what-13-years-in-data-taught-me-about-hype-cycles/</guid>
      <pubDate>Mon, 19 Jan 2026 00:00:00 +0000</pubDate>
      <category>Craft &amp; Communication</category>
      <description>Big data, IoT, digital twins, data lakes, data mesh, the modern data stack, generative AI, agents. Thirteen years and roughly a dozen revolutions in — here&#x27;s the pattern under all of them, and how I&#x27;ve learned to read a hype cycle from the inside.</description>
    </item>
    <item>
      <title>Ignite 2025: Microsoft (Re)discovers Governance</title>
      <link>https://gigilevens.com/writing/ignite-2025-microsoft-discovers-governance/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/ignite-2025-microsoft-discovers-governance/</guid>
      <pubDate>Thu, 27 Nov 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>After two years of agents-everywhere, Ignite 2025 had a different flavour: alongside the &#x27;Frontier Firm&#x27; grandeur, Microsoft spent a surprising amount of energy on governance and control. If you&#x27;ve been arguing that the agent era needs foundations, this was the year the keynote agreed with you.</description>
    </item>
    <item>
      <title>From the Pipeline to the Boardroom: Why the Best Data Advice Comes From People Who&#x27;ve Built It</title>
      <link>https://gigilevens.com/writing/from-the-pipeline-to-the-boardroom-why-the-best-data-advice-comes-from-people-who-built-it/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/from-the-pipeline-to-the-boardroom-why-the-best-data-advice-comes-from-people-who-built-it/</guid>
      <pubDate>Mon, 24 Nov 2025 00:00:00 +0000</pubDate>
      <category>Craft &amp; Communication</category>
      <description>There&#x27;s a whole industry of data advice given by people who&#x27;ve never built a pipeline, met a production incident, or watched a migration go sideways at 2am. I think that&#x27;s exactly backwards — and this is the case for advice that&#x27;s earned at the keyboard first.</description>
    </item>
    <item>
      <title>Microsoft-Deep, Vendor-Neutral: When I Tell Clients It&#x27;s the Wrong Tool</title>
      <link>https://gigilevens.com/writing/microsoft-deep-vendor-neutral-when-i-tell-clients-it-the-wrong-tool/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/microsoft-deep-vendor-neutral-when-i-tell-clients-it-the-wrong-tool/</guid>
      <pubDate>Mon, 22 Sep 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>I&#x27;ve spent my career going deep on the Microsoft data stack. Which is exactly why, when Microsoft isn&#x27;t the right answer for a client, I&#x27;m the person who should say so — and I do.</description>
    </item>
    <item>
      <title>The Em Dash Was Mine First</title>
      <link>https://gigilevens.com/writing/the-em-dash-was-mine-first/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/the-em-dash-was-mine-first/</guid>
      <pubDate>Mon, 11 Aug 2025 00:00:00 +0000</pubDate>
      <category>Craft &amp; Communication</category>
      <description>Someone told me a paragraph I&#x27;d written &#x27;read like ChatGPT&#x27; — because of the em dashes. I&#x27;ve been using them since I was a language student, long before the machines learned to. A short defence of good punctuation in a suspicious age.</description>
    </item>
    <item>
      <title>What Transfers From Building to Advising (and What Doesn&#x27;t)</title>
      <link>https://gigilevens.com/writing/what-transfers-from-building-to-advising-and-what-doesn/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/what-transfers-from-building-to-advising-and-what-doesn/</guid>
      <pubDate>Mon, 16 Jun 2025 00:00:00 +0000</pubDate>
      <category>Craft &amp; Communication</category>
      <description>Moving from building data systems to advising on them isn&#x27;t a promotion to a higher plane — some of what made me good at the first is exactly what I have to keep, and some of it actively gets in the way.</description>
    </item>
    <item>
      <title>Build 2025: Agents Talking to Agents, and a Protocol That Might Matter More</title>
      <link>https://gigilevens.com/writing/build-2025-agents-talking-to-agents-and-a-protocol-that-matters/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/build-2025-agents-talking-to-agents-and-a-protocol-that-matters/</guid>
      <pubDate>Tue, 27 May 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Build 2025 leaned all the way into multi-agent systems — agents delegating to other agents to get work done. The demos were impressive. But the announcement I think actually matters got a fraction of the airtime: a boring-sounding open protocol going GA.</description>
    </item>
    <item>
      <title>Why We&#x27;re Not Migrating Everything to Fabric (Yet)</title>
      <link>https://gigilevens.com/writing/why-not-fully-fabric-yet/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/why-not-fully-fabric-yet/</guid>
      <pubDate>Fri, 23 May 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Fabric is our default for new work — and we&#x27;ve deliberately not moved everything onto it. The discipline of deciding what to keep where, and why &#x27;migrate it all&#x27; is rarely the plan.</description>
    </item>
    <item>
      <title>Copilot in Power BI and Fabric: What&#x27;s Really Useful (and What&#x27;s Just Hype)</title>
      <link>https://gigilevens.com/writing/copilot-usefulness-vs-hype/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/copilot-usefulness-vs-hype/</guid>
      <pubDate>Sat, 17 May 2025 00:00:00 +0000</pubDate>
      <category>Analytics &amp; BI</category>
      <description>After a year and a half of Copilot, an honest map of what&#x27;s genuinely load-bearing, what&#x27;s a nice convenience, and what&#x27;s still theatre — plus the one test that sorts them.</description>
    </item>
    <item>
      <title>The Pitfalls of Fabric Without Governance: Three Horror Stories (and Fixes)</title>
      <link>https://gigilevens.com/writing/fabric-governance-horror-stories/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-governance-horror-stories/</guid>
      <pubDate>Fri, 09 May 2025 00:00:00 +0000</pubDate>
      <category>Data Governance &amp; Compliance</category>
      <description>Three Fabric governance near-disasters — over-shared data, a duplicated &#x27;truth&#x27;, a runaway capacity — and exactly how each was caught and fixed. Learn from these instead of living them.</description>
    </item>
    <item>
      <title>Watchlist: Fabric Features We&#x27;re Monitoring in H2 2025</title>
      <link>https://gigilevens.com/writing/fabric-features-h2-2025-watchlist/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-features-h2-2025-watchlist/</guid>
      <pubDate>Tue, 29 Apr 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>The Fabric features I&#x27;m actually watching for the second half of 2025 — the ones that could genuinely change the work, and the ones I&#x27;ll believe when I see them running.</description>
    </item>
    <item>
      <title>The First 90 Days with Microsoft Fabric: An SME Analytics Leader&#x27;s Journal</title>
      <link>https://gigilevens.com/writing/first-90-days-microsoft-fabric/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/first-90-days-microsoft-fabric/</guid>
      <pubDate>Fri, 18 Apr 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>An honest journal of the first ninety days adopting Microsoft Fabric — the real wins, the surprises the demos never mention, and the bumps worth warning you about before you hit them.</description>
    </item>
    <item>
      <title>Microsoft Sustainability Scorecards: What SMEs Can Track Right Now (and What Not)</title>
      <link>https://gigilevens.com/writing/fabric-sustainability-scorecards-smes/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-sustainability-scorecards-smes/</guid>
      <pubDate>Fri, 11 Apr 2025 00:00:00 +0000</pubDate>
      <category>AI &amp; Responsible Tech</category>
      <description>What Microsoft&#x27;s sustainability data tooling can actually track this spring — energy, cost, estimated emissions — and, just as importantly, the things it can&#x27;t, and you shouldn&#x27;t pretend it does.</description>
    </item>
    <item>
      <title>FabCon 2025: OneLake Security Finally Grows Up (and So Do the Agents)</title>
      <link>https://gigilevens.com/writing/fabcon-2025-onelake-security-finally-grows-up/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabcon-2025-onelake-security-finally-grows-up/</guid>
      <pubDate>Sun, 06 Apr 2025 00:00:00 +0000</pubDate>
      <category>Data Governance &amp; Compliance</category>
      <description>Amid the now-obligatory wall of agent announcements, FabCon 2025 shipped the thing I&#x27;ve quietly wanted for two years: define your security once and have Fabric enforce it across every engine. That&#x27;s the announcement that actually changes how I&#x27;d build.</description>
    </item>
    <item>
      <title>Three Fabric Shortcuts That Changed How We Deliver Reports</title>
      <link>https://gigilevens.com/writing/fabric-shortcuts-reporting-benefits/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-shortcuts-reporting-benefits/</guid>
      <pubDate>Thu, 03 Apr 2025 00:00:00 +0000</pubDate>
      <category>Analytics &amp; BI</category>
      <description>OneLake shortcuts let you use data where it lives instead of copying it everywhere. Three ways that quietly changed how we deliver reports — reuse, trust, and speed.</description>
    </item>
    <item>
      <title>Which Azure Analytics Service Do You *Actually* Need? A Decision Tree for SMEs</title>
      <link>https://gigilevens.com/writing/azure-analytics-decision-tree-sme/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/azure-analytics-decision-tree-sme/</guid>
      <pubDate>Mon, 24 Mar 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Fabric-first doesn&#x27;t mean Fabric-only. A practical decision tree for choosing the right Azure analytics service — including the cases where the honest answer isn&#x27;t Fabric at all.</description>
    </item>
    <item>
      <title>Should SMEs Worry About AI Ethics? A Fabric-Centric View on Responsible Data Use</title>
      <link>https://gigilevens.com/writing/ai-ethics-in-fabric-for-smes/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/ai-ethics-in-fabric-for-smes/</guid>
      <pubDate>Fri, 14 Mar 2025 00:00:00 +0000</pubDate>
      <category>AI &amp; Responsible Tech</category>
      <description>Responsible AI isn&#x27;t a compliance box you tick after the fact. Here&#x27;s what ethical Copilot use actually looks like when it&#x27;s your data and your real decisions on the line.</description>
    </item>
    <item>
      <title>Stop Wasting Data: Fabric Cost Management Through Governance</title>
      <link>https://gigilevens.com/writing/fabric-cost-management-governance/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-cost-management-governance/</guid>
      <pubDate>Wed, 05 Mar 2025 00:00:00 +0000</pubDate>
      <category>Data Governance &amp; Compliance</category>
      <description>The cheapest Fabric capacity is the one you&#x27;re not wasting. Governance isn&#x27;t the opposite of cost control — done properly, it&#x27;s the single most effective form of it.</description>
    </item>
    <item>
      <title>The Case for Retiring Your On-Prem SSIS: A Migration Blueprint to ADF and Fabric Pipelines</title>
      <link>https://gigilevens.com/writing/retire-ssis-adf-fabric-migration/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/retire-ssis-adf-fabric-migration/</guid>
      <pubDate>Sun, 23 Feb 2025 00:00:00 +0000</pubDate>
      <category>Data Engineering</category>
      <description>I spent years building SSIS packages by the hundred. Here&#x27;s my honest blueprint for retiring on-prem SSIS to ADF and Fabric pipelines — and, just as important, what not to carry across.</description>
    </item>
    <item>
      <title>From Power BI Premium to Fabric: How to Rethink Your BI Cost Model</title>
      <link>https://gigilevens.com/writing/powerbi-premium-to-fabric-cost-model/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/powerbi-premium-to-fabric-cost-model/</guid>
      <pubDate>Thu, 13 Feb 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Power BI Premium&#x27;s capacity SKUs are giving way to Fabric&#x27;s. Treat it as a like-for-like swap and you&#x27;ll miss the point — and probably overpay for the privilege.</description>
    </item>
    <item>
      <title>Governance Without Guilt: Making Data Stewardship Realistic for Small Teams</title>
      <link>https://gigilevens.com/writing/governance-without-guilt/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/governance-without-guilt/</guid>
      <pubDate>Thu, 30 Jan 2025 00:00:00 +0000</pubDate>
      <category>Data Governance &amp; Compliance</category>
      <description>Most governance dies because it&#x27;s sold as compliance and red tape. Here&#x27;s how a lean data team makes stewardship light enough that people actually keep doing it.</description>
    </item>
    <item>
      <title>Compliance as a Competitive Edge: The EU AI Act for Boards</title>
      <link>https://gigilevens.com/writing/compliance-as-a-competitive-edge-the-eu-ai-act-for-boards/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/compliance-as-a-competitive-edge-the-eu-ai-act-for-boards/</guid>
      <pubDate>Tue, 21 Jan 2025 00:00:00 +0000</pubDate>
      <category>AI &amp; Responsible Tech</category>
      <description>Most boards are treating the EU AI Act as a cost to minimise and a risk to survive. The ones who&#x27;ll pull ahead are treating it as the opposite — a forcing function that makes them the trustworthy option in a market where trust is about to become scarce and valuable.</description>
    </item>
    <item>
      <title>Fabric Costs Explained: How CUs Work (and What They&#x27;re Worth)</title>
      <link>https://gigilevens.com/writing/fabric-costs-explained-cu-model/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/fabric-costs-explained-cu-model/</guid>
      <pubDate>Fri, 17 Jan 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>What a Capacity Unit actually is, how smoothing and bursting really work, and how to plan Fabric capacity so the invoice never blindsides you or the board.</description>
    </item>
    <item>
      <title>7 Predictions for Microsoft Data &amp; AI in 2025</title>
      <link>https://gigilevens.com/writing/7-predictions-for-microsoft-data-and-ai-in-2025/</link>
      <guid isPermaLink="true">https://gigilevens.com/writing/7-predictions-for-microsoft-data-and-ai-in-2025/</guid>
      <pubDate>Wed, 08 Jan 2025 00:00:00 +0000</pubDate>
      <category>Architecture &amp; Strategy</category>
      <description>Second year of doing this in the open. Seven specific calls on where Microsoft&#x27;s data and AI stack goes in 2025 — agents, Fabric, governance, and the AI Act — dated now so December can mark my homework.</description>
    </item>
  </channel>
</rss>
