{"resourceId":"salisbury-admin-ai-feature-readiness-2026","versions":[{"version":"external-ada9c6791ee3b1392a2b725e278bf21920618c3b2ada28c708c13f976851589c","resource":{"id":"salisbury-admin-ai-feature-readiness-2026","title":"Salisbury documents the work required before enabling embedded administrative AI","organization":"Salisbury University AI Task Force","sector":"Public higher education administration","geography":"Maryland, United States","publishedAt":"Spring 2026; exact publication day unverified","publicationDate":null,"eventDate":null,"sourceName":"AI Task Force Final Report: Operations & Administration","sourceLabel":"Institutional self-assessment and recommendations","sourceUrl":"https://www.salisbury.edu/administration/campus-governance/faculty-senate/_files/25-26/2026-04-14/ai-task-force-rpts/2026-04-14-AI-Task-Force-Fnl-Rpt-Operations-Admin.pdf","evidenceClass":"government-evaluation","outcomeClass":"cautionary","topics":["knowledge-work","data-security","governance-procurement","accessibility-workforce","operating-model"],"finding":"Salisbury's spring self-assessment reports unevaluated embedded AI features and insufficient evaluation capacity.","sledRelevance":"Interpretation: directly useful to campus administrative evaluation; teaching, student-success and research outcomes are outside this item's scope.","evidence":"Workday HR/finance AI features were not enabled pending review. The report identifies missing evaluation procedures and funding; it proposes phased pilots, not completed benefits.","architectureImplications":"Interpretation: trace embedded feature permissions and writes before activation.","governanceImplications":"Interpretation: record the decision owner, evidence threshold and re-review trigger for the selected workflow.","securityPrivacyImplications":"Interpretation: minimize institutional records in tests and verify access boundaries; these sources establish no security effectiveness.","caveats":"Institutional self-assessment, not independent audit. Research appendices remain to be compiled. No baseline, measured savings or evaluation sample is supplied. Spring findings do not establish September status; an internal March follow-up reference prevents inferring a precise publication date from the April file path.","streamIds":["campus-operations"],"roles":{"sales":"Interpretation: the customer problem is a queue of embedded features without a funded route to evaluate them. Ask the CIO, HR, finance, procurement and staff representatives which feature addresses a documented service delay, who can authorize its use, and what evaluation resources exist now. A bounded engagement could inventory one administrative product and design a pilot decision record. The value hypothesis is better prioritization and less unexamined activation, subject to validation. Establish current status before approaching Salisbury or a peer institution. This historical report supports discovery questions; it supplies neither an available sales opportunity nor evidence of savings from any named product.","engineering":"Interpretation: begin with one disabled administrative feature and document the tenant configuration, input fields, permissions, outputs and downstream actions. Require a test environment, a business process baseline and approved synthetic or minimized records. Verify whether the feature only drafts a suggestion or can alter a system of record. Cloud, local and hybrid choices require separate data-path analysis; a license entitlement is insufficient evidence of suitability. Proposed proof of value should compare correctness and review time against the current process, test unauthorized access and demonstrate disablement. The source provides no implementation specification or security test results to reuse as assurance.","delivery":"Interpretation: have an administrative service owner sponsor the pilot, with IT responsible for configuration and HR or finance responsible for business correctness. Dependencies include allocated reviewer time, procurement evidence and a usable training plan. Train staff on exceptions and escalation before introducing the feature into routine work. Require a governance decision before activation and again after material supplier changes. Proposed acceptance is a completed access test, documented owner for each data flow, and a quality-and-time comparison covering representative exceptions. Agree thresholds before testing. Risks include hidden review work, staff uncertainty and allowing a convenient feature to outpace oversight."},"retrievedAt":"2026-09-13T03:00:50Z","enrichedAt":"2026-09-13T03:02:50Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Interpretation: test accessible task completion with affected staff and count training and review effort; no accessibility outcome is established.","procurementImplications":"Interpretation: make pilot continuation depend on evidence, recurring support costs and feasible exit; no supplier performance guarantee follows.","operatingModelImplications":"Interpretation: assign business correctness, technical operation and staff support to named owners before expanding the pilot.","updateExplanation":"New URL in the full 247-resource archive. Fills a specific public-campus embedded-feature activation gap; older evidence, not a new announcement.","sourceVerification":{"openedUrl":"https://www.salisbury.edu/administration/campus-governance/faculty-senate/_files/25-26/2026-04-14/ai-task-force-rpts/2026-04-14-AI-Task-Force-Fnl-Rpt-Operations-Admin.pdf","referenceExcerpt":"AI feature evaluation has not been systematically prioritized.","promptVersion":"sled-research-v3.1","model":null,"basis":"agent-reported inspection"}}}]}