{"resourceId":"und-nebula-ai-pilot-pcard-audit-research-admin-202611","versions":[{"version":"external-fe64c47487322ad5c586dc5463a914a6e770d17cc7b39eb5a9d8be34a65e430d","resource":{"id":"und-nebula-ai-pilot-pcard-audit-research-admin-202611","title":"University of North Dakota opens its Nebula UND AI pilot in November 2026 with a PCard audit and research-award management agent among first projects","organization":"University of North Dakota","sector":"Higher education IT, procurement and research administration","geography":"North Dakota, United States","publishedAt":"Undated pilot page (page metadata last modified 30 October 2026); pilot stated to launch November 2026","publicationDate":null,"eventDate":null,"sourceName":"University of North Dakota University Information Technology","sourceLabel":"University IT pilot page","sourceUrl":"https://campus.und.edu/campus-services/uit/cio/nebula-und-pilot.html","evidenceClass":"standards-guidance","outcomeClass":"emerging","topics":["operating-model","governance-procurement","developers-agents","knowledge-work"],"finding":"UND's IT organisation says its Nebula UND AI pilot launches in November 2026, by invitation or approved project, to test whether the platform delivers value and can be supported at scale. It evaluates capabilities, security and responsible use, user support, and costs and infrastructure. Listed projects include a planned Procurement PCard audit that would organise transaction data, flag missing statements and discrepancies, and direct staff to items needing review while final decisions stay with human reviewers, and an in-progress agent to help manage about 160 active research awards.","sledRelevance":"Why now: the pilot's stated launch month began on 1 November. It is a public-university example of piloting AI in finance controls and research administration with explicit scale-up criteria rather than a campus-wide license.","evidence":"Official pilot page describing evaluation criteria and project goals; no results, metrics or vendor detail.","architectureImplications":"A single campus platform hosts several project agents; PeopleSoft reports and spreadsheets are named as current data sources for Atlas.","governanceImplications":"Participation is gated by invitation or approved project; security and responsible use is a stated evaluation criterion.","securityPrivacyImplications":"PCard transaction data and research award budgets are in scope, so access controls on financial data matter; no controls are described.","caveats":"Pilot design only; PCard project is planned, not built; no publication date, vendor, model or results.","streamIds":["campus-operations"],"roles":{"sales":"Sales framing: UND is piloting AI in procurement controls and research administration before any campus-wide rollout, with value and cost questions written into the pilot. Interpretation: controllers and research administrators at similar public universities may want help scoping one finance-control use case with measurable review targets. A bounded engagement is a PCard or expense audit pilot design with baseline metrics. Do not claim savings or audit improvements; UND lists goals and has not reported results.","engineering":"Engineering framing: Nebula UND is one platform hosting several department agents, with PeopleSoft reports and spreadsheets named as current inputs for research award tracking. Interpretation: pre-sales should ask how the platform reaches PeopleSoft and card-transaction data, what access controls apply to financial records, and how flagged exceptions are logged for audit. The page does not name the vendor, models or hosting, so integration claims need confirmation from UND IT.","delivery":"Delivery framing: UND's pilot explicitly tests whether AI can be supported sustainably and keeps final decisions with human reviewers in the PCard audit. Interpretation: delivery plans should set a baseline for manual review time and exception rates, run AI flagging alongside current audits, and record reviewer overrides. Acceptance should be agreed review-coverage and accuracy targets set before launch, since the page sets none."},"roleDetails":{"sales":{"customerProblem":"Procurement and research administration staff spend time manually reviewing transactions and award data across systems.","stakeholders":["CIO","Procurement & Payment Services director","Controller","Research administration lead"],"discoveryQuestions":["How many PCard transactions are reviewed per cycle, and how long does it take?","Which exceptions matter most to auditors?","What evidence would justify scaling beyond the pilot?"],"valueHypothesis":"AI-assisted exception flagging may reduce repetitive review time; UND has not measured this.","potentialEngagement":"A bounded PCard or expense audit pilot design with baseline and success metrics.","unsupportedClaims":["Audit time savings","Improved fraud or error detection","Scale-up decision or budget"]},"engineering":{"fit":"Campuses piloting department agents on a shared AI platform before broad licensing.","architectureIntegration":"Central platform hosting project agents; inputs include PeopleSoft reports, spreadsheets and transaction data.","prerequisites":["Read access to PeopleSoft and card-transaction data","Role-based access for procurement reviewers","Exception logging"],"constraints":["Vendor, models and hosting not disclosed","Invitation-only pilot scope","PCard project still planned"],"security":"Security and responsible use is an evaluation criterion; specific controls on financial data are not described.","proofOfValue":["Compare AI-flagged exceptions with a completed manual audit","Measure reviewer override rate","Confirm access logs for financial data"]},"delivery":{"work":"Design and run a PCard audit agent alongside the current manual audit.","dependencies":["Procurement & Payment Services","UND IT","Internal audit"],"ownership":"UND Procurement & Payment Services, with UIT as platform owner.","skillsAdoption":"Procurement staff trained to review and override AI flags.","governanceCheckpoints":["Data-access approval for transaction records","Parallel run before reliance","Pilot value and cost review before scaling"],"acceptanceCriteria":["Proposed: AI flags cover all exceptions found by the manual audit in a parallel cycle","Proposed: reviewer time per cycle measured against baseline"],"risks":["Missed exceptions","Over-reliance on flags","Unclear support costs at scale"]}},"retrievedAt":"2026-11-02T08:34:12-06:00","enrichedAt":"2026-11-02T08:38:26-06:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Aims to reduce repetitive manual review for procurement and grant staff and to free staff for higher-value work; no staffing change stated.","procurementImplications":"Pilot is framed to establish costs and infrastructure before broader access; vendor and pricing not disclosed.","operatingModelImplications":"Departments own projects (Procurement & Payment Services, Engineering research support) on a central IT platform.","sourceVerification":{"openedUrl":"https://campus.und.edu/campus-services/uit/cio/nebula-und-pilot.html","referenceExcerpt":"We also want to learn how automation, analytics, and AI can responsibly support university administrative processes while keeping final decisions with human reviewers.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}