Public Records Requests: Two Paths to Getting Ahead of the Flood

The same data set your staff already searches for answers can extend to a resident-facing portal for common questions, like a short-term rental fee or the permit history on a parcel. A five-year review of one mid-sized city's requests found that community development and parcel questions made up 40% to 60% of total volume, with contracts and purchase orders close behind. Not every data set belongs in a public portal. Each agency decides what's exposed and what stays behind a formal request.

For requests that do need a full review, the AI assistant searches Outlook, SharePoint, Laserfiche, Granicus, permitting systems, and your ERP in one pass. It pre-reviews records for responsiveness, threads and dedupes email so staff see conversations instead of repeated fragments, and flags PII, PHI, and privileged content for redaction. Records move through a clear pipeline: non-responsive, withheld, needs review, cleared by AI, and ready for release. A staff member always makes the final call to release a record. Permissions scale with your team, so a subject matter expert can review a record without seeing every file, and the system routes and tracks each step for audit purposes.

Clovis received a six-part request spanning multiple departments and projects. Scoping it the traditional way would have taken weeks of huddles to figure out who to search and what timeline to cover. Using Clovis's citywide Madison index, staff configured the search and had records ready for review in four and a half hours. Deduping and threading also brought the initial 7,600 records down to about 1,800 that were genuinely responsive, so the team reviewed a fraction of what the search first surfaced.