Launching with the first pilot cohort
Evaluate municipal AI on evidence you can put in the file.
Elunia is an evaluation workspace for public authorities. It benchmarks candidate AI systems against real municipal workflows, records the evidence behind every result and assembles a decision pack an officer can sign.
- Evaluation surface
- Workflow-level
- Permits, intake, records
- Evidence
- Digest-anchored
- Prompt, config, output
- Outcome
- Decision pack
- Conditions and sign-off
EVAL-2026-0148 · sample
Evidence previewBuilding permit completeness triage — candidate B
Proceed with conditions· Evaluation panelSystem prompt digest
Register for the pilot
Join the first cohort of authorities evaluating AI assistance with Elunia. We contact you before a cohort is confirmed — no obligation, no procurement commitment.
What an Elunia evaluation covers
Four commitments shape every run. Each one leaves a record a reviewer can inspect.
Evidence, not impressions
Every score is anchored to a stored prompt, model configuration and output digest, so a reviewer can re-open the exact material months later.
Municipal workflow benchmarks
Candidate systems run against representative caseloads — permits, social support intake, records requests — instead of generic leaderboards.
Statutory conditions
Each requirement is tracked as compliant, conditional, non-compliant or insufficient evidence, with the reference that produced the judgement.
Human decision, recorded
Elunia recommends and documents. Officers adjudicate, sign off and keep an audit trail of who decided what, and on which evidence.
How the pilot runs
Four stages from scoping to a signed decision pack.
Questions before you register
Bring one workflow. Leave with a defensible record.
Pilot places are limited per cohort so each authority gets scoping time with an evaluation lead.