{"resourceId":"kktv-cspd-axon-sarah-nonemergency-20260922","versions":[{"version":"external-175ee34ab28e163d2caf1f73c25a1efee4d0f508f67b6efd582be6a899a41ad2","resource":{"id":"kktv-cspd-axon-sarah-nonemergency-20260922","title":"Colorado Springs Communications Center launches Axon Prepared AI; S.A.R.A.H. answers non-emergency calls first while humans keep all 911","organization":"Colorado Springs Police Department Public Safety Communications Center","sector":"Local emergency communications / PSAP","geography":"Colorado Springs, Colorado, United States","publishedAt":"2026-09-22","publicationDate":"2026-09-22","eventDate":"2026-09-22","sourceName":"KKTV","sourceLabel":"Local TV news report on CSPD communications-center AI launch","sourceUrl":"https://www.kktv.com/2026/09/23/colorado-springs-police-department-launches-ai-system-enhance-emergency-non-emergency-call-response/","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["operating-model","accessibility-workforce","infrastructure","governance-procurement"],"finding":"On 2026-09-22 CSPD announced its Public Safety Communications Center has launched Prepared by Axon after April 2026 testing. Trained call takers still answer all 911 calls; the AI agent S.A.R.A.H. now answers non-emergency calls first. Non-emergency traffic is more than half of telephone volume. The stack transcribes calls, translates in dozens of languages, identifies critical details, generates summaries, and helps route non-emergency calls to police and fire. Callers may still request a person; staff will review performance and update responses as community needs change. No independent answer-time or diversion metrics are published in the article.","sledRelevance":"Direct PSAP/dispatch operating-model evidence for emergency-services: joint communications-center AI that protects human 911 answering while offloading non-emergency load that includes fire and medical service requests. Transferable to other multi-agency PSAPs evaluating vendor AI front doors. Policing brand on the agency is noted; this edition uses the piece for dispatch/fire/medical routing and 911 answer-time intent, not for investigative policing use cases.","evidence":"KKTV quotes CSPD that Prepared by Axon assists emergency and non-emergency response; testing began April 2026; call takers continue to answer all 911; S.A.R.A.H. answers non-emergency first; non-emergency is over half of calls; transcription, real-time translation in dozens of languages, critical-detail identification, summaries, and routing to police and fire; human option retained; regular performance review planned. No measured before/after KPIs in the article.","architectureImplications":"Expect a telephony/IVR or voice-agent front door integrated with CAD/call-taking workflows, real-time ASR/translation, and summarization that updates call records. Keep a hard path for all 911 to human telecommunicators and a clear escalate-to-human control for non-emergency sessions. Plan language coverage, summary quality monitoring, and failover if the AI agent is unavailable.","governanceImplications":"Document the non-emergency-only scope, human 911 retention, caller notice about AI, and change-control for response content as services change. Align review cadence with communications-center leadership and elected oversight. Avoid marketing autonomous emergency triage from this launch announcement alone.","securityPrivacyImplications":"Call audio, transcripts, translations, and AI summaries are sensitive public-safety data. Require vendor data-handling terms, retention limits, access controls, and audit of who can read summaries. Clarify whether model training uses local call content.","caveats":"Police-department press framing; independent measured outcomes (answer-time deltas, diversion rate, false-route rate) are not in the article. Axon vendor capabilities described via agency release. Medical routing is mentioned at system-purpose level without EMS outcome data.","streamIds":["emergency-services"],"roles":{"sales":"Frame Colorado Springs as a non-emergency AI front door that keeps humans on every 911 call—not autonomous emergency triage. Discovery should quantify today’s non-emergency share, language demand, fire/medical routing accuracy, and baseline 911 answer-time KPIs before promising relief. Bounded engagement: PSAP readiness assessment and pilot design for Axon-class non-emergency intake with mandatory human escalate, caller human option, and published review metrics. Unsupported claims include guaranteed 911 answer-time gains, safe unattended life-threat handling, or EMS clinical improvement from this launch announcement alone.","engineering":"Fit is a voice-agent plus transcription, translation, and summary layer ahead of CAD for non-emergency trunks, while 911 remains human-answered end to end. Integrate carefully with telephony, recording, and CAD wrap-up fields; define hard failover when S.A.R.A.H. is down so queues do not black-hole. Prerequisites include call-flow maps, language packs, and a clear human-takeover UX for telecommunicators. Constraints include joint police/fire misroute risk and Axon platform lock-in. Proof of value should measure diversion rate, escalate accuracy, translation complaints, and 911 answer-within-15-seconds—not vendor feature lists.","delivery":"Delivery work includes call-flow redesign, agent content governance, telecommunicator takeover drills, and ongoing updates as city services change. Dependencies span Axon Prepared configuration, telephony/CAD vendors, and supervisor capacity for weekly AI-session QA. Ownership stays with the Public Safety Communications Center, with fire/EMS liaisons reviewing routing accuracy. Skills adoption needs trust that AI is not punitive scoring. Governance checkpoints: go-live proving human 911 path, 30-day metrics review, and change-control for AI responses. Acceptance requires logged human 911 answering, tested caller human-option, and published diversion metrics."},"roleDetails":{"sales":{"customerProblem":"PSAPs need to shed non-emergency load so telecommunicators can protect 911 answer times without replacing humans on emergencies.","stakeholders":["Communications center directors","Fire/EMS dispatch supervisors","City IT/security","Elected officials watching 911 metrics"],"discoveryQuestions":["What share of calls is non-emergency today?","Which languages must be covered?","How is escalate-to-human logged and audited?","Which 911 answer-time KPI is the success metric?"],"valueHypothesis":"An AI non-emergency front door can free capacity for 911 if humans retain all emergency calls and diversion quality is measured.","potentialEngagement":"Bounded PSAP intake pilot design and KPI workshop for Axon-class non-emergency agents—not a guaranteed citywide AI 911 program.","unsupportedClaims":["Guaranteed faster 911 answers from this launch alone","Safe AI handling of life-threat calls","Proven EMS outcome improvement"]},"engineering":{"fit":"Voice AI agent on non-emergency trunks with ASR/translation/summary feeding CAD, while 911 stays human-answered.","architectureIntegration":"Telephony front door to AI agent to CAD/RMS notes; parallel recording and human takeover path; monitoring for agent outage failover to humans.","prerequisites":["Mapped non-emergency vs 911 call flows","CAD API or wrap-up fields for AI summaries","Language pack requirements","Recording/retention policy alignment"],"constraints":["911 must remain human-answered per agency claim","Joint police/fire routing increases misroute risk","Vendor lock-in to Axon Prepared stack"],"security":"Protect transcripts/summaries as CJIS/HIPAA-adjacent sensitive data; define training-data use and access logging.","proofOfValue":["Measure non-emergency diversion and escalate rates","Track 911 answer-within-15s before/after","Sample translation/summary error reviews weekly"]},"delivery":{"work":"Configure S.A.R.A.H. flows, train call takers on takeover, publish caller guidance, and stand up performance review rituals.","dependencies":["Axon Prepared licensing","Telephony/CAD vendors","Communications center supervisors","City service directory accuracy"],"ownership":"Public Safety Communications Center owns operations; IT owns integrations; fire/EMS liaisons own routing accuracy feedback.","skillsAdoption":"Telecommunicators need takeover drills and trust that AI is not scoring them punitively; supervisors need QA sampling of AI sessions.","governanceCheckpoints":["Go-live checklist confirming human 911 path","30-day metrics review","Content change-control board for AI responses"],"acceptanceCriteria":["100% of 911 calls still human-answered in logs","Caller human-option path tested","Published diversion and escalate metrics for first 30 days"],"risks":["Misrouting fire/medical non-emergency requests","Public trust backlash if AI answers unexpected emergencies","Unmeasured vendor claims overstated in procurement"]}},"retrievedAt":"2026-09-22T22:17:00-05:00","enrichedAt":"2026-09-22T22:17:00-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Translation in dozens of languages may improve access for LEP callers on non-emergency lines; workforce benefit is claimed capacity for 911, but burnout reduction is not measured here.","procurementImplications":"Sole-source or platform-aligned Axon procurement is likely; require measurable SLAs for diversion quality, uptime, and human escalate—not feature marketing alone.","operatingModelImplications":"Shifts non-emergency intake to AI-first with human backup; emergency path unchanged. Requires ongoing content stewardship as city services change.","sourceVerification":{"openedUrl":"https://www.kktv.com/2026/09/23/colorado-springs-police-department-launches-ai-system-enhance-emergency-non-emergency-call-response/","referenceExcerpt":"non-emergency calls will now first be answered by an artificial intelligence agent known as S.A.R.A.H.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}