SLED AI Adoption Intelligence

https://sled.lighthouseadvisory.consulting/feeds/microsoft.xmlSLED AI Adoption Intelligence · Microsoft2026-11-11T05:16:05.347ZLighthouse Advisoryhttps://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-10Microsoft · Issue 12 · 2026-11-102026-11-11T05:16:05.347Z2026-11-10T00:00:00.000Z

Four items. Ohio State and Microsoft announced a five-year partnership on Nov. 10 to move about 70-80% of the medical center's storage and compute to Azure, build a Fabric data foundation and expand Dragon Copilot note drafting (vendor claims; one self-reported facility metric, no financial terms). Microsoft's Nov. 10 security release included a critical Azure OpenAI flaw, CVE-2026-119990 (CVSS 9.8, no authentication required); ZDI notes the guidance changed on release day to 'no customer action needed', and it is not reported as exploited. Portland's Microsoft 365 Copilot privacy impact assessment, which lists 48 risks and excludes police use, reaches its stated validity end today, and Australia's DTA guidance (via IDM) warns that Copilot outputs may not be reproducible for records purposes. No new U.S. public-safety, HHS, transportation/utility or K-12 Microsoft AI evidence surfaced. 4 sources · 1 cross-source patterns Government Copilot reviews place records capture and periodic reassessment on the deploying agency, not the platform When does your Copilot privacy or records review expire, and does it cover how relied-upon outputs are captured as records?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-09Microsoft · Issue 11 · 2026-11-092026-11-10T11:25:14.534Z2026-11-09T00:00:00.000Z

Six items, mostly higher education and cross-SLED licensing. CRN reports that on Nov. 9 Copilot Chat fair-use caps reached organizations above 2,000 seats; Microsoft's support page says they apply to work or school accounts without a Copilot add-on, and admins cannot see or adjust them (vendor policy, no usage data). George Mason's updated rollout page reports 412 Premium license holders and 99% utilization for Oct. 1-28 under role-based allocation; Manchester (UK) set full student Microsoft 365 Copilot for 2027/28 after a pilot. Directions on Microsoft gives independent per-tier education pricing, including the end of volume discounts from May 1, 2027. Microsoft says a student lock-down to its Study and Learn tutor is coming (vendor claim). Colorado State's Nov. 9 staff AI hub story was dropped on resubmit: the same URL was already published in campus-operations the same day with no substantive new facts for an updateExplanation. No new U.S. public-safety, HHS, infrastructure or research-enterprise Microsoft AI evidence surfaced. 6 sources · 2 cross-source patterns Universities are rationing paid Microsoft 365 Copilot by role or readiness gates rather than licensing everyone Which roles or programmes will get paid Copilot first, who funds them, and what evidence triggers expansion? The no-cost Copilot tier is being constrained as Microsoft moves full AI into paid tiers If Copilot Chat becomes intermittently unavailable to heavy users, what is the approved escalation path, and is it funded?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-07Microsoft · Issue 10 · 2026-11-072026-11-08T05:04:14.209Z2026-11-07T00:00:00.000Z

Three new items. The Seattle Times reported Nov. 7 that Seattle Public Schools closed a Gemini-via-search loophole and keeps Microsoft Copilot staff-only, yet told its board it cannot turn off AI rewrite features in Word, as parents rallied for blocking all K-12 generative AI including features in approved tools. Microsoft's Nov. 5 Microsoft 365 Education portfolio announcement (vendor claim) positions A7 as its most advanced AI tier and retires Office 365 A1 and the Student Use Benefit on Jan. 1, 2028, with pricing changes from May 1, 2027. The School Data Leadership Association's analysis explains A0's limits, the loss of the 100 TB storage base and the difference between Copilot Chat and A7's Microsoft 365 Copilot. No new US public-safety, HHS, infrastructure or research-enterprise Microsoft AI evidence surfaced. 3 sources · 1 cross-source patterns K-12 AI exposure is being decided by Microsoft suite licensing and embedded features, not just standalone chatbot policy For each student and staff group, which Microsoft AI features does our licence mix enable, which can we disable, and are those decisions reviewed before our next renewal?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-06Microsoft · Issue 09 · 2026-11-062026-11-07T05:47:40.184Z2026-11-06T00:00:00.000Z

Three new items. Broward County Public Schools named its first Microsoft AI/Copilot Innovation Award winners on Nov. 5, publishing school-level use cases and more than 5,300 Copilot interactions since April at Pompano Beach High under a Microsoft-funded $4,000 recognition program (district reporting; no independent learning outcomes). Computerworld's Nov. 6 scrutiny says Microsoft is extending undisclosed Copilot Chat fair-use caps for non-Premium enterprise users (Nov. 2 for ≤2,000 seats; Nov. 9 for larger), that admins cannot view or override limits, and that analysts warn of shadow AI and budgeting risk. Ingeniøren reports Odense Municipality issued 1,000 Copilot licences, offers Edge Copilot to other admin staff, and firewall-blocked ChatGPT, Claude and Mistral—while noting 100 FTEs freed since 2022 were not primarily from AI. No new US public-safety, HHS, infrastructure or higher-ed Microsoft measured evaluation surfaced. 3 sources · 1 cross-source patterns Public-sector Copilot scale-ups are pairing Microsoft seats with controls against consumer AI—and publishing usage or seat counts When we standardize on Microsoft Copilot, what firewall, approved-tool and measurement controls keep citizen/student data out of consumer AI—and which efficiency claims are actually Copilot-attributable?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-05Microsoft · Issue 08 · 2026-11-052026-11-06T15:18:57.064Z2026-11-05T00:00:00.000Z

Three new items. Washington State University warned that Microsoft will expand unpublished fair-use limits on Copilot Chat features for non-Premium users on Nov. 9, and administrators cannot view or adjust those limits. The Register's Nov. 5 report corroborates the commercial rollout (Nov. 2 for smaller tenants, Nov. 9 for larger) and the upgrade-prompt path when caps are hit. BusinessWorld reports Microsoft, DepEd and DICT launching AI-literacy credentials for 2,000 Filipino public-school teachers on top of a claimed Copilot access grant for 1 million educators (vendor claim via independent report; Philippines transferability caveat). UNCG's Nov. 5 AI Showcase story was dropped on resubmit: the same URL was already published in campus-operations the same day with no substantive new facts for an updateExplanation. No new US public-safety, HHS, infrastructure or independent measured Microsoft AI evaluation surfaced. 3 sources · 1 cross-source patterns Included Copilot Chat capacity is becoming an opaque, vendor-controlled fair-use limit for non-Premium users How will we detect, explain and govern Copilot Chat fair-use limit hits without admin-visible quotas, and what evidence—not just a throttle—justifies Premium upgrades?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-11-04Microsoft · Issue 07 · 2026-11-042026-11-05T05:43:56.679Z2026-11-04T00:00:00.000Z

Two new items. Microsoft says GPT-5.6, Edit with Copilot in Word, Excel and PowerPoint, and Copilot in Forms are now in GCC, GCC High and DoD, narrowing the government-cloud gap that has shaped agency Copilot decisions (vendor claim, no performance data). UW-Madison opened department purchase of Copilot Premium at $201.36 per licence per year, approved for FERPA but not HIPAA data, with mandatory training. No new independent evaluation or audit of Microsoft AI in SLED surfaced; public safety, HHS, infrastructure and K-12 lacked new evidence.

A one-item edition. MassHousing, an independent Massachusetts state agency, issued an RFP on Nov. 2 for a consultant to roll Microsoft Copilot out to all staff in FY27. The agency already named Copilot its official AI platform and runs a pilot; its GCC tenant on G5 licences includes Copilot Basic but not Premium, and the consultant must judge from data-readiness reporting whether Premium is warranted. Responses are due Nov. 16. No new independent evaluation, audit or measured outcome for Microsoft AI in SLED surfaced since the last run; gaps remain in public safety, HHS, infrastructure, higher education and K-12.

A thin two-item edition for Nov. 1-2. Elk Point, Alberta, adopted an AI policy that makes Microsoft Copilot and approved Microsoft apps the only sanctioned tools, bans consumer ChatGPT and Gemini for town work, and still bars personal, confidential, financial, legal and HR data from any AI system, with human review and records treatment required. Hamilton-Wentworth District School Board confirmed that Copilot goes live Nov. 3 for Grades 9-12 only, starting in Study and Learn mode, with alerts to board staff on inappropriate or safety-related use; it promises to assess impact but publishes no measures. Both are Canadian, so US transferability is limited to policy design. No new US, HHS, public-safety, infrastructure or higher-ed Microsoft evidence and no independent measured outcomes were found.

A thin, governance-heavy edition covering Oct. 30 to Nov. 1. Two Washington local governments show how public-records law is shaping Copilot use. Chelan's council got a first look on Oct. 27 at a draft AI policy (reported Oct. 31). It makes Microsoft Copilot the preferred tool, with free Copilot only for public information and enterprise Copilot for city data, reserves an enterprise Claude for complex work, and treats prompts and outputs as potential public records. King County's October records guidance names Copilot Chat as the approved tool. It warns that its 30-day storage setting is not legal retention, so substantive uses must be captured into county records by hand. Separately, a Microsoft blog sets a forced migration: SharePoint Agents move to Copilot Agent Builder from mid-November, legacy .agent files stop working after March 2027, and Copilot Chat-only users may incur Copilot Credit charges. No independent measured outcomes were found, and there were no new Microsoft items for HHS, public safety, transportation, K-12 or higher education in the window.

A follow-up edition about allocation and accountability more than launches. Dallas's council approved its $25.4M Microsoft Government Enterprise renewal on Oct. 28. The city's record calls the figure a ceiling that encumbers nothing, against about $8.9M in average annual spend, so any Copilot add-on would be funded department by department. In higher education, the University of Hawaiʻi gave every student and employee free Microsoft 365 with Copilot Chat after a systemwide agreement cut costs about 14%. UVic opened paid Copilot only by departmental application, after permission reviews and labelling. The City of London Corporation is piloting Copilot Premium for members with a six-month reassign-if-unused rule. Canada's federal IT agency says it cannot track Microsoft licence usage. In K-12, Microsoft's default teen protections depend on correct identity data, the AFT standard opens Nov. 1 amid Fairplay's enforcement critique, and Seattle faces scrutiny over free Copilot through Microsoft Elevate. Researchers also report an unpatched VS Code issue. No independent measured outcomes were found. 9 sources · 3 cross-source patterns Copilot Chat is becoming the default for everyone, while paid Copilot is rationed through applications, pilots and readiness gates What readiness checks and value evidence must a department show before it gets paid Copilot seats beyond the Copilot Chat baseline? Licence ceilings and idle seats make utilisation reporting a precondition for AI add-on spend Can we report Copilot licence activity by department, and who reclaims seats that go unused? K-12 Microsoft AI protections are arriving through defaults and contracts, while independent critics question student-facing use itself Before enabling student Copilot, have we fixed identity data, adopted a policy and defined the learning outcomes we will measure?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-10-30Microsoft · Issue 02 · 2026-10-302026-10-31T04:41:01.551Z2026-10-30T00:00:00.000Z

A second Microsoft edition in which adoption stays bounded while the commercial model shifts. Delaware opened Copilot Chat in its government tenant to participating agencies after mandatory training. Dallas's $25.4M three-year Microsoft renewal through Texas DIR permits Copilot or other AI add-ons, with 46% of 14,000 licences in public safety. Middlesex County (Ontario) starts with a leadership Copilot pilot. UBC sells premium Copilot to faculty and staff but withholds agents, and Tacoma launched Copilot for grades 8-12 after a required lesson. On the commercial side, Microsoft's Foundry Copilot harness starts metered billing Dec. 1, and new standalone Copilot offers open to education and government that day with default usage billing for Copilot Business in CSP. Microsoft's own documentation shows hosted agents unavailable in Azure Government. Directions on Microsoft warns of lock-in and unpredictable spend. In education, Microsoft's AFT-agreed protections open to all US districts Nov. 1, McGraw Hill plans a Copilot plugin, and Slovakia ran out of Copilot teacher licences. No independent measured outcomes were found. 13 sources · 4 cross-source patterns Public institutions are opening Microsoft Copilot in bounded tiers, gated by training, role, grade or funding, and holding back agents What gate (training, role, grade, funding) and what baseline measure will decide when our Copilot access expands, and what must happen before agents are allowed? Microsoft AI is moving from seats to meters, and cost governance is becoming a core control Who owns AI consumption budgets and caps in our Microsoft estate, and how will we compare spend with measured value before Dec. 1 billing changes take effect? New Microsoft agent capabilities launch in commercial Azure first; Azure Government lags, and preview features may lack compliance commitments Which agent capabilities in our design are generally available, with compliance commitments, in the cloud our data must live in? In education, Microsoft AI is arriving with contract protections, content integrations and allocation programmes, but no outcome evidence Beyond contract terms and access, what learning or teaching outcome will we measure to judge Microsoft AI in our schools?https://sled.lighthouseadvisory.consulting/streams/microsoft/editions/2026-10-29Microsoft · Issue 01 · 2026-10-292026-10-30T03:22:23.816Z2026-10-29T00:00:00.000Z

A first Microsoft edition focused on government-cloud timing and scoped adoption. Microsoft says Microsoft 365 G7 and Agent 365 are purchasable for eligible GCC customers, with capabilities phased by authorization and more purchasing options from Nov. 1. Directions on Microsoft counters that G7 reached only GCC, lacks many E7 AI features and, with Agent 365, some audit, DLP and runtime controls, and advises minimal investment for now. SAMexpert finds A7's student benefit excludes Copilot and no A7/G7 prices are published. Planet's Oct. 29 digest shows Edit with Copilot in Excel reaching government clouds. Adoption remains scoped: Guilford County's 50-licence GCC pilot (no stated success measures), HWDSB's Grades 9-12 Copilot rollout from Nov. 3 and George Mason's department-funded Azure AI Foundry model. Microsoft's own AI-literacy posts report reach, not outcomes. No new independent measured evaluation of Microsoft AI in SLED was found. 10 sources · 3 cross-source patterns Government-cloud Copilot and agent capabilities arrive in phases, and independent analysts document gaps behind the launch messaging Which of the Copilot and agent capabilities we are paying for, or being asked to buy, are live and governable in our GCC tenant today, and which are roadmap promises? Institutions gate Microsoft AI access by scope (licence count, grade band, data class and chargeback) rather than enterprise-wide, often without published success measures For our scoped Copilot or Foundry rollout, what baseline and success measures will decide whether we expand, change or stop? Microsoft's education and library AI-literacy programmes report reach and self-reported perception, not measured outcomes If we adopt Microsoft's free AI-literacy materials, how will we measure whether staff and students actually gain verification and responsible-use skills?