{"resourceId":"la-vergne-aurelian-ava-nonemergency-2026","versions":[{"version":"external-ca2724a92a83aac3c64d5aeded58544aa1f9c0edc7f0011489d4e427dc94e83a","resource":{"id":"la-vergne-aurelian-ava-nonemergency-2026","title":"La Vergne deploys Aurelian AVA on non-emergency line amid 73% non-emergency call share","organization":"City of La Vergne Emergency Communications Center (Aurelian AVA)","sector":"911 / PSAP emergency communications (police and fire dispatch)","geography":"La Vergne, Tennessee, United States","publishedAt":"2026-09-16","publicationDate":"2026-09-16","eventDate":"2026-09-16","sourceName":"Rutherford Source","sourceLabel":"Independent local news reporting","sourceUrl":"https://rutherfordsource.com/la-vergne-launches-virtual-assistant-to-handle-non-emergency-calls/","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["operating-model","accessibility-workforce","governance-procurement","infrastructure"],"finding":"On 2026-09-16 Rutherford Source reported La Vergne’s Emergency Communications Center deployed Aurelian’s AVA on the non-emergency line to gather information, create CAD calls for service, and transfer potential emergencies to telecommunicators, citing ~50,000 annual calls with non-emergency about 73% (33,367) versus 12,299 911—while vendor claims of 60–75% automation and ~15 hours/week savings are not presented as La Vergne audited outcomes.","sledRelevance":"Why now: published on the edition date after the 2026-09-15 last run; extends the archived Jonesboro/Maine/Weber non-emergency AI pattern with a same-day U.S. fire/EMS-serving PSAP deployment announcement and explicit emergency-transfer control.","evidence":"Article states deployment, CAD create capability, emergency monitoring/transfer, language support (55+), Speck understaffing quote, volume split, and Aurelian typical partner-agency automation/hours claims. No La Vergne misroute or diversion audit table.","architectureImplications":"Non-emergency conversational agent integrated to CAD for call-for-service creation, with emergency-intent detection transferring to human telecommunicators; multi-channel (voice/text/web) alongside traditional trunks.","governanceImplications":"Publish scope (non-emergency only), require escalation QA before treating vendor 60–75% claims as local performance, and brief served fire/police agencies on AI-created CAD incidents.","securityPrivacyImplications":"Voice/text/web AI sessions carry caller identity and location—apply PSAP retention, vendor training limits, and access controls comparable to recorded non-emergency lines.","caveats":"Announcement journalism; vendor metrics not local audits; center also serves police; no hang-up or critical-escalation rates published.","streamIds":["emergency-services"],"roles":{"sales":"Frame AVA as non-emergency capacity relief for understaffed PSAPs that still transfer potential emergencies to humans—not as a 911 call taker or proven FTE replacement. Discovery should separate La Vergne’s volume split from Aurelian’s partner-agency 60–75% automation claim and ask what local diversion, misroute, and hang-up metrics will be published in 90 days. Offer a bounded pilot with CAD QA sampling and public notice rather than a staffing-cut business case. Unsupported claims include guaranteed 15 hours/week saved per telecommunicator, safe AI answering of 911, or clinical outcome gains from this Rutherford Source deployment story alone.","engineering":"Fit is a multi-channel non-emergency virtual assistant that can create CAD calls for service and hand off when emergency cues appear. Architecture needs hard separation of 911 trunks, reliable emergency-intent escalation, CAD identity for AI-originated incidents, and session logging for QA. Prerequisites include intent taxonomy, language/channel test plans, and vendor limits on training use of call audio. Proof of value is audited diversion share, critical-escalation success, hang-up rate, and telecommunicator available time—not vendor typical 60–75% figures.","delivery":"Stand up non-emergency AVA, train reclaim and CAD review of AI-created incidents, publish caller-facing scope, and sample sessions before expanding features or overnight reliance. Ownership stays with the Emergency Communications Manager; QA owns stop-rules; PIO owns public explanation. SkillsAdoption keeps emergency call-taking sharp while supervising AI. Governance checkpoints include 90-day local metrics review versus vendor claims. Risks include treating vendor averages as local SLAs, mission creep onto 911, and under-reviewed CAD entries created by AI."},"roleDetails":{"sales":{"customerProblem":"Non-emergency volume (~73% here) competes with 911 and radio for scarce telecommunicators, while leaders refuse unattended AI on true emergencies.","stakeholders":["Emergency Communications Manager","Telecommunicators","Fire Rescue and police chiefs served by the center","City leadership / PIO"],"discoveryQuestions":["What local diversion and misroute metrics will be published at 30/90 days?","How are AI-created CAD incidents labeled and reviewed?","What is the emergency-transfer success rate on tested scripts?","How is AI disclosed to non-emergency callers?"],"valueHypothesis":"If AVA reliably absorbs routine non-emergency work and escalates emergently, telecommunicators may regain attention for 911—subject to local audited diversion, not vendor partner averages.","potentialEngagement":"90-day non-emergency pilot design with CAD QA sampling, public-notice draft, and stop-rules tied to escalation failures.","unsupportedClaims":["Guaranteed 60–75% local automation","15 hours/week saved per telecommunicator as proven here","Safe AI answering of 911","Staffing reductions justified by this announcement"]},"engineering":{"fit":"Non-emergency multi-channel virtual assistant with CAD write and emergency transfer—not live 911 answering.","architectureIntegration":"Non-emergency DNIS/channels → AVA dialog → CAD CFS create or human queue on emergency cues; recording/QA store aligned with PSAP retention.","prerequisites":["Emergency-phrase escalation tests","CAD field for AI-originated incidents","Vendor DPA / training-use limits","Staff reclaim playbooks"],"constraints":["No local audited diversion yet","Vendor metrics are partner generalizations","Policing and fire share the same center capacity"],"security":"Encrypt recordings/transcripts; restrict vendor model training on caller audio; monitor social-engineering against the agent.","proofOfValue":["Local diversion % on audited logs","Critical escalation success rate","Hang-up / abandoned rate versus baseline","Telecommunicator available minutes during peaks"]},"delivery":{"work":"Configure AVA, train reclaim and CAD review, publish scope, and run a 90-day metrics gate before expansion.","dependencies":["Aurelian/CAD admin access","QA auditor time","Legal/PIO review","Minimum emergency staffing floor"],"ownership":"Emergency Communications Manager owns scope; QA owns sampling/stop-rules; PIO owns public messaging.","skillsAdoption":"Keep emergency call-taking proficiency while teaching AI supervision and CAD QA of AI-created incidents.","governanceCheckpoints":["Written non-emergency-only policy","Public notice approval","90-day local metrics versus vendor claims"],"acceptanceCriteria":["Documented emergency-transfer tests passed","AI-originated CAD incidents reviewable in logs","Published description of AI scope"],"risks":["Vendor claim treated as SLA","Mission creep onto 911","Under-reviewed AI CAD entries"]}},"retrievedAt":"2026-09-17T03:08:30Z","enrichedAt":"2026-09-17T03:11:57Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"55+ language and text/web channels may help some callers, but human fallback must remain for complex or emergency situations; staffing shortages remain the root driver.","procurementImplications":"Separate vendor typical performance claims from contractual SLAs; require local diversion/escalation reporting and exit rights at 90 days before multi-year lock-in.","operatingModelImplications":"Operate AVA as a non-emergency capacity valve with humans owning emergencies and CAD QA; do not reduce minimum 911 staffing based on unvalidated automation percentages.","sourceVerification":{"openedUrl":"https://rutherfordsource.com/la-vergne-launches-virtual-assistant-to-handle-non-emergency-calls/","referenceExcerpt":"AVA takes the routine calls off our plates so our telecommunicators can give their full attention","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}