{"resourceId":"brandeis-suny-morrisville-ai-help-desk-agents-edtech-20261023","versions":[{"version":"external-55c67dd34c7465da19a2d8fb584bb43a5f835c0f0c9aa59919e294bf2c652e50","resource":{"id":"brandeis-suny-morrisville-ai-help-desk-agents-edtech-20261023","title":"Brandeis reports 36% chat and 32% voice containment six weeks into AI help desk agents, after a three-month knowledge base rebuild and a no-FERPA-data scope","organization":"Brandeis University; SUNY Morrisville (via EdTech Magazine)","sector":"Private and public universities: IT service desk","geography":"Waltham, Massachusetts and Morrisville, New York, United States","publishedAt":"23 October 2026 (EdTech Magazine: Focus on Higher Education)","publicationDate":"2026-10-23","eventDate":"2026-10-23","sourceName":"EdTech Magazine (CDW)","sourceLabel":"Trade publication feature","sourceUrl":"https://edtechmagazine.com/higher/article/2026/10/how-ai-agents-are-helping-campus-it-teams-manage-student-support-during-peak-demand-perfcon","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["developers-agents","operating-model","data-security","knowledge-work"],"finding":"SUNY Morrisville reports a 35% decrease in tickets routed to help desk agents since deploying Wilson, a Microsoft Teams assistant that answers common questions and creates trackable tickets. Brandeis phased in AI chat (July, faculty and staff) and voice (August) assistants, reaching students at move-in. Six weeks in, chat containment is 36% and voice 32%, inside the 30% to 50% goal set for the first six months; voicebot call volume matched last year while need for human technicians dropped by more than half. The build took over eight months, anchored by a three-month knowledge base revamp that found conflicting documentation. Scope was deliberately limited to Tier 0 and Tier 1 topics with no ERP, CRM or student information system integrations.","sledRelevance":"Gives early, quantified operator results for AI service desk agents during peak demand and shows the preparation (knowledge base cleanup, narrow data scope) behind them. Useful benchmark for campuses weighing similar deployments.","evidence":"Feature article quoting institution officials' own figures; no independent evaluation, baseline period or definition of containment is provided.","architectureImplications":"Retrieval-augmented voice and chat agents over an enterprise knowledge base; Teams-hosted assistant with ticket creation at Morrisville.","governanceImplications":"Brandeis limited initial scope to topics that need no access to FERPA-protected systems; future knowledge base changes are to be coordinated with the AI assistants.","securityPrivacyImplications":"Only student data available to the Brandeis assistant is a username a student volunteers to create a ticket; authentication-related support workflows still put agents near identity systems.","caveats":"Six-week figures; self-reported; publisher is a reseller with commercial interest in help desk AI.","streamIds":["campus-operations"],"roles":{"sales":"Sales framing: Brandeis and SUNY Morrisville give early, concrete help desk numbers: 36% chat and 32% voice containment at six weeks, and a 35% drop in routed tickets. Interpretation: the strongest offer is readiness, since Brandeis says its three-month knowledge base rebuild was the most underestimated part. A bounded engagement is a knowledge base audit and Tier 0 scope design. Avoid promising staffing cuts; Brandeis says reducing staff was not the goal.","engineering":"Engineering framing: these are retrieval-based agents whose accuracy depends on clean documentation, and Brandeis kept them away from any system holding FERPA data. Interpretation: pre-sales should inventory conflicting knowledge sources, define which ticket categories are in scope, and confirm ticket creation works without student-system integration. Because the agents handle password and MFA questions, review how they hand off to identity workflows without being able to change credentials themselves.","delivery":"Delivery framing: Brandeis took more than eight months, phased staff before students, and set a six-month containment target of 30% to 50%. Interpretation: delivery plans should budget documentation cleanup first, publish a target and its definition before launch, and report containment by category and channel. Assign an owner who keeps the knowledge base and the assistants in sync, since Brandeis ties future knowledge base changes to the AI assistants."},"roleDetails":{"sales":{"customerProblem":"IT service desks face seasonal ticket spikes and want round-the-clock answers to routine questions without touching protected student data.","stakeholders":["CIO","IT service desk director","Knowledge management owner","Information security and FERPA compliance"],"discoveryQuestions":["How consistent is the current knowledge base?","Which ticket categories make up most peak volume?","What containment target would count as success?","Which systems must the agent never touch?"],"valueHypothesis":"Interpretation: well-prepared Tier 0 agents may contain roughly a third of routine contacts early on; figures are self-reported.","potentialEngagement":"A bounded knowledge base audit and Tier 0 agent scope design for one service desk.","unsupportedClaims":["Staff reductions","Containment beyond six weeks","Vendor-specific results","Guaranteed savings"]},"engineering":{"fit":"Fits campuses with high-volume, low-complexity IT requests and documented answers.","architectureIntegration":"Retrieval-augmented chat and voice agents over the knowledge base, with ticket creation; Teams-hosted at Morrisville.","prerequisites":["Cleaned, de-duplicated knowledge base","Ticketing integration","Defined Tier 0 and Tier 1 scope","Voice channel integration"],"constraints":["No ERP, CRM or student information system integration in Brandeis's current scope","FERPA limits","Containment definitions vary"],"security":"Brandeis agent sees only a volunteered username; identity-related questions require careful handoff design.","proofOfValue":["Measure containment by ticket category against a defined baseline","Test answer accuracy after knowledge base edits","Verify the agent cannot reach FERPA systems"]},"delivery":{"work":"Deploy and run AI chat and voice agents for the IT service desk in phases.","dependencies":["Knowledge base revamp","Ticketing system","Communications for move-in and peak periods"],"ownership":"IT client services, with a named knowledge base owner.","skillsAdoption":"Technicians move to complex support; documentation authors need training on writing for agents.","governanceCheckpoints":["Scope review before any student-data integration","Containment target and definition agreed before launch","Knowledge base change control that includes agents"],"acceptanceCriteria":["Proposed: containment within a stated target range by six months","Proposed: escalation rate and repeat-contact rate reported monthly","Proposed: zero agent access to FERPA systems verified quarterly"],"risks":["Conflicting documentation producing wrong answers","Overstated savings","Scope creep into protected data"]}},"retrievedAt":"2026-10-23T21:48:42-05:00","enrichedAt":"2026-10-23T21:52:27-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Technicians shift toward complex and high-touch support; round-the-clock availability is the stated aim rather than staff reduction.","procurementImplications":"Vendors are not named for Brandeis; CDW publishes the article and quotes its own architect, so it should be weighed as partly promotional.","operatingModelImplications":"Documentation quality becomes an operational dependency; knowledge base governance must include the AI assistants.","sourceVerification":{"openedUrl":"https://edtechmagazine.com/higher/article/2026/10/how-ai-agents-are-helping-campus-it-teams-manage-student-support-during-peak-demand-perfcon","referenceExcerpt":"Six weeks into deployment at Brandeis, the chat assistant is currently hitting a 36% containment rate, while the voice assistant is at 32%.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}