{"resourceId":"wyandotte-gx-agent-treasury-callcut-20260914","versions":[{"version":"external-e8b8a9b96fd84f615c698d6a328c6a0ed96d19922a02829c1e03dee8a623e5a8","resource":{"id":"wyandotte-gx-agent-treasury-callcut-20260914","title":"Wyandotte County/Kansas City, KS reports 30% Treasury call-volume cut from AI Government Experience Agent","organization":"Unified Government of Wyandotte County and Kansas City, Kansas (via Route Fifty)","sector":"County/city unified government resident services","geography":"Wyandotte County / Kansas City, Kansas, USA","publishedAt":"September 14, 2026","publicationDate":"2026-09-14","eventDate":null,"sourceName":"Route Fifty","sourceLabel":"Independent reporting / interview","sourceUrl":"https://www.route-fifty.com/people/2026/09/were-all-just-trying-help-local-service-citizen-winner-says/415962/","evidenceClass":"independent-reporting","outcomeClass":"effective","topics":["operating-model","accessibility-workforce","knowledge-work","governance-procurement"],"finding":"Route Fifty’s Sept. 14, 2026 interview with Crystal Sprague (director of innovation and performance) states the unified government handles ~310,000 calls yearly and piloted a multilingual, around-the-clock AI Government Experience Agent in Treasury that led to a 30% reduction in call volume, freeing staff for more meaningful work, with plans to scale across other agencies. Sprague stresses careful, non-hype AI deployment, ADA checks on AI-produced materials, and measuring quality of interactions—not just call counts. The effort contributed to a Service to the Citizen award.","sledRelevance":"Direct local-government resident-services outcome: a mid-sized unified city/county publishing a call-deflection KPI for an AI CX agent, with explicit caution on change management and accessibility—highly transferable to other 311/Treasury front doors.","evidence":"On-record interview with the jurisdiction’s innovation director; metric is operator-reported (30% Treasury call-volume reduction) without an independent audit in the article. Contextual figures: ~310k annual calls; innovation team ~7 FTEs. Vendor ecosystem (Granicus) discussed as future enterprise relationship, not as the measured study design.","architectureImplications":"Resident-facing multilingual digital assistant integrated with CX/311-style channels; knowledge grounding and escalation to staff; future agent layer across communications products requires identity, content, and logging controls.","governanceImplications":"Scale only with checks/balances, elected literacy on AI types, and public-service quality measures (e.g., callback likelihood) alongside deflection; avoid technology-for-its-own-sake purchases.","securityPrivacyImplications":"Minimize sensitive data in AI interactions; govern what information is published for model consumption; apply same quality standards (including ADA) as human-produced materials.","caveats":"Single-jurisdiction interview; 30% figure lacks published methodology/time window; vendor success stories exist separately and were not used as the evidence base. Scaling claims are prospective.","streamIds":["local-government"],"roles":{"sales":"Treasury and 311 leaders need AI that returns staff time without replacing complex counter work. Route Fifty’s Sept. 14 interview with Wyandotte/Kansas City, Kansas innovation director Crystal Sprague reports a Government Experience Agent piloted in Treasury that cut call volume 30%, with plans to scale across agencies after a Service to the Citizen award. Frame discovery on baseline call mix, multilingual coverage, and what questions stay human-only. Bound value to published call-volume and interaction-quality gains—not enterprise AI transformation or guaranteed FTE cuts from one department pilot.","engineering":"Fit is a resident-facing multilingual digital assistant for routine Treasury and similar inquiries, not case adjudication. Architecture likely sits on a CX/311 platform (Granicus ecosystem referenced) with knowledge grounding and escalation to staff. Prerequisites: approved content corpus, call-volume baseline, privacy rules for chat logs, and ADA-quality review for AI outputs. Constraints: small innovation team (~7 FTEs), taxpayer scrutiny of shiny buys, and change-management readiness across departments. Prove value by re-measuring call volume, average handle time for complex calls, and callback rates—not prompt counts.","delivery":"Delivery work is pilot-to-scale: expand beyond Treasury with department change management, content ownership, and outcome dashboards. Dependencies include vendor enterprise agreements, 311 operations insight, and elected-official literacy on agentic vs generative tools. Ownership: innovation/performance with Treasury/311 operational owners. Skills: staff who escalate and quality-check AI answers. Governance checkpoints: data-entry bans for sensitive info, ADA review of AI-produced materials, and stop criteria if accuracy or resident satisfaction slips. Risks: scaling before content is curated; treating call deflection as success without longer quality interactions."},"roleDetails":{"sales":{"customerProblem":"Routine Treasury and resident calls consume staff time needed for complex counter work amid rising inquiry volume (~310k calls/year cited).","stakeholders":["Director of innovation/performance","Treasury manager","311 leadership","County commissioners/mayor","CX platform vendor"],"discoveryQuestions":["What share of Treasury calls are routine hours/scheduling vs complex cases?","What multilingual and after-hours coverage is required?","Which escalation paths and accuracy checks exist before scale?","How will call-volume and callback-quality KPIs be published?"],"valueHypothesis":"A governed multilingual CX agent can cut routine call volume (~30% in Treasury pilot) and return staff to higher-value resident interactions if content and escalation are solid.","potentialEngagement":"Six-week CX-agent readiness: baseline call taxonomy, content gaps, privacy/ADA checks, and a scale/no-scale scorecard beyond Treasury.","unsupportedClaims":["Guaranteed FTE reductions countywide","That 30% call cut generalizes to every department without re-baselining","Autonomous resolution of complex financial or legal resident cases"]},"engineering":{"fit":"Strong for routine multilingual FAQ and service navigation; weak for adjudicated Treasury disputes.","architectureIntegration":"Website/311/CX layer with grounded answers, human escalation, analytics on deflection and failure modes; optional voice channel later.","prerequisites":["Approved knowledge corpus","Call baseline metrics","Privacy/retention rules","ADA review process for AI outputs"],"constraints":["Small innovation team capacity","Departmental change readiness","Taxpayer scrutiny of platform spend"],"security":"Minimize PII in prompts/logs; tenant or vendor controls; no training on resident data without agreement; access logging.","proofOfValue":["Treasury call volume vs baseline","Average handle time on remaining complex calls","Callback/repeat-contact rate","Accuracy sample of AI answers"]},"delivery":{"work":"Expand pilot department-by-department with content owners, training, and KPI dashboards; maintain 311 feedback loop.","dependencies":["Vendor enterprise relationship","Treasury/311 SMEs","Commission communication on AI limits"],"ownership":"Innovation/performance owns program; departments own content accuracy and escalations.","skillsAdoption":"Staff learn when to trust, edit, or override AI answers; leaders avoid celebrating raw AI usage.","governanceCheckpoints":["Pre-scale accuracy review","Sensitive-data entry ban","ADA check on AI-produced materials","Public explanation of AI role"],"acceptanceCriteria":["≥ documented call-volume improvement in pilot dept","Named human escalation path live","No unresolved critical accuracy defects in sample audit","Scale plan with stop criteria"],"risks":["Scaling on award narrative without re-measurement","Content drift","Staff fear vs hype polarizing adoption"]}},"retrievedAt":"2026-09-26T22:18:37-05:00","enrichedAt":"2026-09-26T22:18:37-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Multilingual 24/7 access can help non-English speakers and after-hours callers; ADA compliance for AI-generated flyers/notices must still be human-checked as Sprague notes.","procurementImplications":"Prefer enterprise CX agreements with measurable deflection/quality SLAs and data-use limits over one-off department chatbot buys.","operatingModelImplications":"Shifts routine FAQ load to AI so Treasury staff can lengthen complex interactions; requires content ownership and 311 insight into failure modes.","sourceVerification":{"openedUrl":"https://www.route-fifty.com/people/2026/09/were-all-just-trying-help-local-service-citizen-winner-says/415962/","referenceExcerpt":"The assistant was first piloted in the government’s Treasury Department, and led to a 30% reduction in call volume.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}