{"resourceId":"clariti-honolulu-civcheck-q1-2026-metrics","versions":[{"version":"external-ba87f355adefbac139a4e1f706c6794a877385bfa4f1fde1ef61bf1555408a65","resource":{"id":"clariti-honolulu-civcheck-q1-2026-metrics","title":"Clariti case study: Honolulu CivCheck cuts residential plan-review cycles 58% and days-to-decision 55% (Q1 2026 sample)","organization":"Clariti (CivCheck) / City and County of Honolulu Department of Planning and Permitting","sector":"Municipal planning and building permitting","geography":"Honolulu, Hawaii, United States","publishedAt":"Undated vendor case study page (metrics labeled Q1 2026; related city award coverage July 2026)","publicationDate":"2026-03-31","eventDate":null,"sourceName":"Clariti","sourceLabel":"Vendor case study with city director quotes and Q1 2026 performance table","sourceUrl":"https://www.claritisoftware.com/honolulu-cuts-review-times-by-70-percent-with-civchecks-ai-plan-review-software","evidenceClass":"vendor-claim","outcomeClass":"mixed","topics":["operating-model","governance-procurement","knowledge-work","data-security"],"finding":"Clariti reports Honolulu DPP's CivCheck guided AI plan review reduced average plan-review cycles from 3.4 to 1.4 (−58%), corrections from 23.5 to 7.7 (−67%), and days to permit decision from 73 to 32.5 (−55%), with ~40.5 applicant days saved versus non-CivCheck applications. Figures compare 19 completed CivCheck residential permits to 17 non-CivCheck projects in Q1 2026; an earlier pilot on 145 pre-screened applications claimed ~70% residential plan-review time reduction. Director quotes emphasize applicant guidance before staff review, not autonomous approval.","sledRelevance":"Direct local-government permitting modernization with published cycle-time metrics—highly transferable as a measurement template, with caution on small completed-sample size and vendor authorship.","evidence":"Vendor-published metrics table and methodology note (19 vs 17 completed residential permits, Q1 2026); pilot claim of 70% review-time cut across 145 pre-screened applications; attributed quotes from DPP Director Dawn Takeuchi Apuna.","architectureImplications":"Applicant-facing guided checks in front of human plan review; standardize rule packs; instrument correction cycles and calendar time as first-class telemetry.","governanceImplications":"Treat vendor ROI slides as hypotheses until locally replicated; disclose sample sizes; keep human final decisions; stage mandatory-use only after measured pilots.","securityPrivacyImplications":"Plan sets are sensitive; contracts should bar training on applicant documents and define retention aligned to permit records.","caveats":"Vendor-authored page; completed-sample n=19/17 is small; selection effects possible between CivCheck and non-CivCheck cohorts; Honolulu mandate coverage already in archive via other URLs—this page newly archives the detailed Q1 table.","streamIds":["local-government"],"roles":{"sales":"Sales framing: Honolulu's CivCheck case argues guided AI plan review can cut correction cycles and calendar time when applicants fix issues before staff review. Use the published Q1 2026 comparison of 19 CivCheck versus 17 non-CivCheck residential permits as a proof conversation with development-services directors—not a guaranteed ROI model for every city. Ask how many completed permits underpin any vendor slide, whether cohorts are comparable, and whether human reviewers still own final decisions before discussing mandates.","engineering":"Engineering framing: Fit is pre-submission completeness and code-check guidance integrated with intake, not autonomous permit issuance. Architecture typically pairs applicant upload, rule-based AI checks, and reviewer queues; prerequisites include codified residential rules and baseline cycle metrics the city already trusts. Constraints include small completed-sample sizes and jurisdiction-specific codes; security should bar vendor training on plan sets and keep public-records retention schedules intact for uploaded drawings.","delivery":"Delivery framing: Stand up a time-boxed pilot with paired CivCheck and control cohorts, publish correction-cycle and days-to-decision KPIs, and train intake staff on exception handling when the tool misfires. Ownership stays with planning and permitting leadership, with IT security co-owning data controls. Acceptance criteria should restate the vendor's claimed reductions against the city's own baseline before any mandatory-use expansion to additional project types."},"roleDetails":{"sales":{"customerProblem":"Residential permit backlogs driven by incomplete or non-compliant submissions force repeated review cycles and long applicant waits.","stakeholders":["Planning & Permitting director","Building officials/plan reviewers","CIO/enterprise architect","Mayor/CA housing lead","Applicant/industry liaison"],"discoveryQuestions":["What are current average review cycles and days-to-decision for residential permits?","How many completed permits will be in any before/after evaluation cohort?","Who owns final approval after AI flags?","Will use be voluntary or mandatory for eligible residential types?"],"valueHypothesis":"Guided AI pre-check can reduce correction cycles and calendar time if applicants remediate before staff review, but published gains rest on small Q1 2026 samples.","potentialEngagement":"Eight-week discovery: baseline metrics extract, code-rule coverage map for eligible residential types, and a pilot evaluation design with paired control permits.","unsupportedClaims":["Guaranteed 70% review-time cut for every jurisdiction","Autonomous code-compliant permit issuance without human review","That n≈19 completed permits generalizes citywide performance"]},"engineering":{"fit":"Strong for applicant-facing guided completeness/code-issue flagging ahead of human plan review; weak for discretionary zoning judgments.","architectureIntegration":"Applicant portal → CivCheck guided checks → corrected package → DPP intake/plan-review systems; optional analytics on cycle counts.","prerequisites":["Digitized residential submittal path","Codified check rules for eligible project types","Baseline permit performance data","Staff exception workflow"],"constraints":["Jurisdiction-specific building codes","Small early completed-sample sizes","Mandatory-use change management","Public-records handling of plan sets"],"security":"Encrypt plan uploads; contractually prohibit vendor training on applicant documents; role-based reviewer access; retain logs for audit.","proofOfValue":["Replicate cycle/correction/days-to-decision metrics on local completed permits","Compare intake rejects before vs after guided checks","Measure staff hours per residential permit","Track applicant remediation completion before first review"]},"delivery":{"work":"Configure residential rule packs, run a measured pilot with control cohort, train intake/review staff, and decide voluntary vs mandatory rollout gates.","dependencies":["Vendor configuration support","DPP process owners","Legal/public-records review","Applicant communications"],"ownership":"Department of Planning and Permitting director with IT security co-ownership of data controls.","skillsAdoption":"Train reviewers to treat AI flags as triage aids; train applicants/industry on remediation workflow; keep human final review.","governanceCheckpoints":["Pilot evaluation report before mandate","Security/privacy contract review","Public metrics dashboard approval","Expansion criteria for commercial types"],"acceptanceCriteria":["Statistically reviewed local reduction in average plan-review cycles vs baseline","Documented human-final-decision policy","No unauthorized vendor reuse of plan data","Published limitations including sample size"],"risks":["Overgeneralizing vendor small-n results","Applicant pushback on mandatory AI tool","Rule drift as codes update","Equity impacts if atypical projects are poorly handled"]}},"retrievedAt":"2026-10-03T22:18:00-05:00","enrichedAt":"2026-10-03T22:25:32.481-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Aims to reduce rework for applicants and free reviewers for complex work; equity risk if atypical projects are poorly handled by rule packs.","procurementImplications":"Demand local baseline replication, sample-size disclosure, and data-use prohibitions in contracts before sole-source or mandatory deployment.","operatingModelImplications":"Shifts incompleteness discovery earlier to applicants; staff become exception handlers over first-line completeness checkers.","sourceVerification":{"openedUrl":"https://www.claritisoftware.com/honolulu-cuts-review-times-by-70-percent-with-civchecks-ai-plan-review-software","referenceExcerpt":"Plan review cycles decreased 58%, from an average of 3.4 cycles to 1.4 cycles.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}