{"resourceId":"ohio-dys-ai-youth-call-monitoring-transcription-rfp-5pct-wer-20261022","versions":[{"version":"external-38ae6ef545ea9c39acda6965bd19bbf3bfcd3f4c67aa824efb1481f75d0b44cf","resource":{"id":"ohio-dys-ai-youth-call-monitoring-transcription-rfp-5pct-wer-20261022","title":"Ohio youth-corrections RFP seeks AI monitoring and transcription of youth communications, with a 5% word-error ceiling and quarterly audits","organization":"ODYS (Ohio Department of Youth Services, as named in the listing)","sector":"Corrections (juvenile) / AI communications monitoring","geography":"Ohio, United States","publishedAt":"RFP posted Oct. 22, 2026; proposals due Nov. 18, 2026","publicationDate":"2026-10-22","eventDate":null,"sourceName":"RFPPlanet (procurement listing)","sourceLabel":"Transcription and Monitoring Investigating Tool (SYS-7027)","sourceUrl":"https://rfpplanet.com/sys-7027/transcription-and-monitoring-investigating","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["data-security","governance-procurement","knowledge-work"],"finding":"A procurement listing for solicitation SYS-7027, posted Oct. 22 with proposals due Nov. 18, 2026, seeks offender communication monitoring and transcription services for ODYS. Requirements include compatibility with the current youth communication provider ViaPath, AI analysis of calls and other digital assets with near-real-time flags, “hot lists” of gangs and other targets, semantic analysis in English, Spanish and Arabic, speaker differentiation, AI summaries, prompt-based querying, link analysis and continuous learning of slang. Transcription must have a word error rate of 5% or less, audited quarterly by the contractor with results provided to the agency. The contractor must also supply analysts to support the agency and law enforcement partners.","sledRelevance":"One of the few corrections AI procurements to write a measurable accuracy requirement and audit cadence into the specification; relevant to jails and youth facilities buying similar call-monitoring tools.","evidence":"Opened the RFPPlanet listing and read the full requirements text displayed.","architectureImplications":"Must integrate with the incumbent inmate communication system and ingest historical communications for AI analysis.","governanceImplications":"Quarterly transcription audits with results to the agency; agency defines hot lists that focus analysis.","securityPrivacyImplications":"Monitoring of youth communications; data handling terms are not shown in the listing.","caveats":"Third-party aggregator listing; the original solicitation documents were not opened, and the agency is identified only by the ODYS acronym. Requirements are not results. Monitoring of juveniles raises heightened rights and privacy questions not addressed in the listing.","streamIds":["public-safety"],"roles":{"sales":"Sales framing: this RFP shows a corrections buyer writing measurable accuracy and audit terms into an AI monitoring purchase. Facilities buying similar tools may want help defining requirements and evaluation. A bounded engagement could be an RFP requirements and evaluation-criteria review for communications analytics. The listing cannot support claims about tool accuracy, investigative value or safety outcomes, and youth monitoring carries rights questions the listing does not address.","engineering":"Engineering framing: the solution must integrate with ViaPath and ingest historical communications, transcribe at 5% word error rate or better in English and support Spanish and Arabic analysis. Pre-sales should ask how word error rate is measured (dialects, slang, call audio quality) and how flags are explained. A proof of value could score transcription on a sample of real calls by language and report error rates separately for each.","delivery":"Delivery framing: work would include integration with the incumbent call platform, hot-list governance, alert routing, quarterly transcription audits and analyst support. Dependencies include ViaPath cooperation and agency rules on who may create hot lists. The operational owner is the agency’s investigations unit. Acceptance should be measured by audited error rates and reviewed alerts, with risks around over-flagging, language bias and youth privacy."},"roleDetails":{"sales":{"customerProblem":"Youth and adult corrections agencies cannot listen to call volume manually and want AI flags, but need accuracy and oversight terms.","stakeholders":["Corrections investigators","Facility administrators","Procurement officers","Legal and youth advocates","Incumbent call provider"],"discoveryQuestions":["How will word error rate be measured across languages and call quality?","Who can add a person or group to a hot list?","How are AI flags reviewed before action?"],"valueHypothesis":"Measured accuracy terms and audits could make AI call monitoring more accountable than manual sampling alone.","potentialEngagement":"Requirements and evaluation-criteria review for a corrections communications analytics procurement.","unsupportedClaims":["Accuracy of any bidder’s tool","Investigative or safety outcomes","Compliance with youth privacy law"]},"engineering":{"fit":"Corrections agencies with an incumbent inmate communication platform.","architectureIntegration":"Ingests call and digital communications from ViaPath; transcription, translation, semantic flags, link analysis and alerts via web portal, email or text.","prerequisites":["Incumbent platform integration","Historical data access","Language coverage"],"constraints":["Juvenile population","Aggregator listing only","Multilingual accuracy varies"],"security":"Restrict hot-list creation and alert access; log queries; define retention for transcripts and summaries.","proofOfValue":["Proposed: measure word error rate on sample calls per language","Proposed: review a sample of flags for false positives before go-live"]},"delivery":{"work":"Integrate with the call platform, configure hot lists and alerts, run quarterly transcription audits and manage analyst support.","dependencies":["ViaPath integration","Agency hot-list policy","Contractor analysts"],"ownership":"Agency investigations unit","skillsAdoption":"Train investigators to treat AI flags as leads and verify against recordings.","governanceCheckpoints":["Hot-list approval process","Quarterly audit review","Annual privacy review"],"acceptanceCriteria":["Proposed: audited word error rate at or below 5%","Proposed: documented human review for every flag acted on"],"risks":["Over-flagging","Language and dialect bias","Privacy harm to youth and the people they call"]}},"retrievedAt":"2026-10-25T02:57:46Z","enrichedAt":"2026-10-25T03:02:27Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Multilingual analysis (Spanish, Arabic) and contractor-provided analysts may supplement agency staff.","procurementImplications":"Specifies a 5% word error rate ceiling and audit duty, a useful contract model; also bundles analyst services.","operatingModelImplications":"Shifts call review from manual listening to AI-flagged review with contractor analyst support.","sourceVerification":{"openedUrl":"https://rfpplanet.com/sys-7027/transcription-and-monitoring-investigating","referenceExcerpt":"Any transcription service used by the vendor as part of the platform shall have an error rate of 5% or less","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}