{"resourceId":"wdgph-ai-chatbot-community-support-91-percent-resolution-20261104","versions":[{"version":"external-1f69dc31e10ad7b52a28fc2be429b37bb6520c4a305dce946f684c81d2b59539","resource":{"id":"wdgph-ai-chatbot-community-support-91-percent-resolution-20261104","title":"Wellington-Dufferin-Guelph Public Health reports AI chatbot averaging 91% automated resolution and 37% after-hours use","organization":"Wellington-Dufferin-Guelph Public Health (Ontario)","sector":"Local public health contact center / community navigation","geography":"Wellington, Dufferin and Guelph, Ontario, Canada","publishedAt":"2026-11-04","publicationDate":"2026-11-04","eventDate":"2026-11-04","sourceName":"Wellington-Dufferin-Guelph Public Health Board of Health report BH.01.NOV0426.C12","sourceLabel":"Local public health agency board report (primary)","sourceUrl":"https://www.wdgpublichealth.ca/sites/default/files/2026-10/BH.01.NOV0426.C12_Client_Community_Support_Q1-Q2.pdf","evidenceClass":"government-evaluation","outcomeClass":"effective","topics":["knowledge-work","accessibility-workforce","operating-model"],"finding":"In its Nov. 4, 2026 Board of Health report on January–June 2026 operations, WDG Public Health describes an AI-powered chatbot that answers open-ended questions using the agency website and staff-curated reputable sources. On average each month the chatbot was opened 196 times with 138 engaged conversations; the agency reports a 91% automated resolution rate and notes that 37% of conversations occur on weekends and after hours. Highest-volume topics were vaccine records, STI testing appointments and vaccination scheduling. Staff continue coaching response quality, and optional age/residence questions are planned to support QI without identifying users.","sledRelevance":"A concrete local public-health chatbot deployment with published volume, after-hours share and resolution metrics—useful benchmark for US local health departments designing 24/7 navigation AI with human coaching and privacy-preserving analytics.","evidence":"Primary agency board report with operational metrics for Q1–Q2 2026. Resolution rate is agency-defined accuracy/relevance, not an external audit.","architectureImplications":"Interpretation: generative AI over website plus curated knowledge base; dashboard under development with Data and Analytics. Models/vendors not named.","governanceImplications":"Staff coaching of responses and planned optional demographics for QI show an iterative governance loop; crisis-escalation detail is not described.","securityPrivacyImplications":"Optional, broad age and area questions intended to protect anonymity; retention of chat content not detailed.","caveats":"Canadian local public health; self-reported 91% resolution without published external validation or safety incident rates.","streamIds":["health-human-services"],"roles":{"sales":"Sales framing: local public health buyers want measured after-hours deflection with staff still coaching answers—not an unsupervised medical chatbot. Use WDGPH's 196 opens / 138 conversations / 91% resolution / 37% after-hours as a conversation starter for US LHDs, while noting the metric is self-defined. Do not claim clinical outcomes; topics are records and appointments.","engineering":"Engineering framing: retrieval-grounded generative answers over agency web content plus a curated knowledge base, with a QI dashboard for topics, resolution and content gaps. Engineering priorities are knowledge-base freshness, after-hours escalation paths for clinical or crisis intents, and privacy-preserving analytics that avoid retaining identifiable chat content while still supporting continuous improvement.","delivery":"Delivery framing: launch with public-health nurse coaching of answers, publish a simple engagement dashboard, and add optional non-identifying demographics only after privacy review. Track whether the after-hours conversation share stays high without raising escalation failures, and schedule recurring content refresh for vaccine and sexual-health topics that dominate volume."},"roleDetails":{"sales":{"customerProblem":"Public health contact centers need 24/7 answers for high-volume topics without abandoning live PHN coverage.","stakeholders":["Medical officer of health / LHD director","Contact-center and PHN supervisors","Data/analytics leads"],"discoveryQuestions":["How do you define and audit automated resolution?","What is the after-hours escalation path?","Which topics are allowed in the knowledge base?"],"valueHypothesis":"Bounded: grounded chatbots can resolve routine navigation after hours if staff coach quality; clinical safety is unproven here.","potentialEngagement":"LHD chatbot pilot design with resolution and escalation metrics.","unsupportedClaims":["Clinical advice quality","External validity of the 91% rate","Call-center FTE reductions"]},"engineering":{"fit":"Public-health website-grounded generative chatbot for navigation and appointment topics.","architectureIntegration":"Chat channel over CMS/website content and curated knowledge base, with analytics dashboard.","prerequisites":["Curated approved content","Escalation to live staff","Topic and resolution telemetry"],"constraints":["Unnamed model/vendor","Self-defined resolution metric","Sensitive sexual-health queries"],"security":"Minimize chat retention, avoid collecting identifiers, and keep optional demographics coarse.","proofOfValue":["Engaged conversation rate","After-hours share","Sampled answer accuracy by PHNs"]},"delivery":{"work":"Content curation, soft launch, PHN coaching cadence, dashboard build, privacy review of demographics.","dependencies":["CCS staffing for coaching","Web content owners","Privacy office"],"ownership":"Community Health operations with Data and Analytics.","skillsAdoption":"Train PHNs to review transcripts and refine prompts/content.","governanceCheckpoints":["Pre-launch content approval","Monthly resolution and topic review","Privacy sign-off on demographics"],"acceptanceCriteria":["Proposed: ≥90% sampled answers rated accurate by PHNs","Proposed: documented escalation for clinical/crisis intents"],"risks":["Incorrect health guidance","Privacy regression if demographics expand","Channel that outpaces content maintenance"]}},"retrievedAt":"2026-11-05T22:09:47-06:00","enrichedAt":"2026-11-05T22:16:10-06:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Extends access when live answer is unavailable; reduces after-hours voicemail load for PHNs and reception.","procurementImplications":"Vendor and contract not disclosed in the report.","operatingModelImplications":"Chatbot as parallel channel to phone/email/in-person CCS, with continuous dashboard-driven improvement rather than one-time launch.","sourceVerification":{"openedUrl":"https://www.wdgpublichealth.ca/sites/default/files/2026-10/BH.01.NOV0426.C12_Client_Community_Support_Q1-Q2.pdf","referenceExcerpt":"The automated resolution rate remains consistently high at an average of 91%.","promptVersion":"sled-research-v4.0 HHS nightly brief","model":null,"basis":"agent-reported inspection"}}}]}