Lighthouse AdvisorySLED AI Adoption Intelligence
← Back to results

From the Local Government edition of September 12, 2026

Academic researchCautionaryNewly relevant · Jun 2025

Atlanta 311 study identifies gaps in request interpretation and case visibility

Georgia Institute of Technology researchers · Municipal resident services · Atlanta, Georgia, United States

Publisher
arXiv / DIS 2025 paper
Original publication
June 13, 2025, arXiv version 1
Source retrieved
2026-09-13
Read original source

What happened

Residents encountered interpretation problems and insufficient visibility into request handling.

Why it matters

Historical U.S. municipal evidence newly added to test resident access beyond answer accuracy; it does not describe current Atlanta performance.

Evidence and measured results

Researchers analyzed 144 survey comments from January 2023–April 2024 and 16 interviews using task walkthroughs and thematic analysis. Tables describe participants and design concerns. The case concerns a keyword-based chatbot; proposed LLM improvements were not tested.

Limitations and uncertainty

Small, single-city qualitative sample skewed toward civic engagement. No causal benefit estimate or evaluation of a replacement LLM.

Put this evidence to work

Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-09-13; this does not change the original publication date. Labels below come from the analysis itself.

Sales

Role takeaway

A 311 director, public works manager, CIO and resident-access lead can investigate repeat contacts and opaque case status. Ask whether a resident can correct an address, identify the responsible agency and challenge an unresolved closure. A bounded engagement could trace one common request category across channels. The value hypothesis is fewer avoidable repeat contacts and clearer accountability. Do not promise call deflection or infer current failure rates from this historical qualitative case. For a smaller municipality, first check whether it has enough staff and records-system access to operate an additional channel.

Pre-sales engineering

Role takeaway

Fit is strongest for an assistant backed by reliable service records. Map address validation, category selection, identity checks, CRM creation and work-order updates. Use approved synthetic cases to test jurisdiction boundaries, incomplete descriptions, duplicate reports and stale status. If adding an LLM, compare it with the existing interface rather than assuming it resolves the observed problems. Limit write permissions during the trial and exercise human handoff. Proposed validation should measure successful task completion and recovery across channels, including screen-reader use. The study supplies no cloud, on-premises or hybrid performance comparison.

Delivery

Role takeaway

The 311 operations owner should jointly pilot the workflow with a receiving department, records staff and accessibility specialists. Establish baseline completion and repeat-contact measures, train agents to recover failed submissions, and review resident feedback before expansion. Dependencies include authoritative status data and staff empowered to reopen cases.

Proposed acceptance
every test request reaches its intended owner, status matches the work-order record, and failed intake can be recovered without lost information. Set a local task-completion target before testing. These are future criteria. Risks include improving conversational polish while unresolved service work remains invisible.

Implementation considerations

Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.

Architecture and integration

What must connect, and where does the AI sit in the workflow?

Validate the connection between conversational intake, jurisdiction lookup, CRM records and work-order status before adding agent actions.

Governance

Who approves, reviews and stays accountable for outcomes?

Give residents a way to contest closure and identify the accountable service team.

Security and privacy

What data, permissions and controls need testing?

Test authorization for case-status access; avoid exposing individual reports when aggregating neighborhood issues.

Accessibility and workforce

Who is affected, and what skills or accommodations follow?

Recruit less engaged residents and people needing language or assistive support; retain staffed alternatives.

Procurement

What should contracts, pricing and exit terms secure?

Require demonstrations of completed resident tasks and usable case handoff in acceptance testing.

Operating model

Which teams own the service once it runs?

Assign ownership of unresolved requests separately from ownership of the chatbot platform.

What changed

Absent from all 247 archive resources checked at offsets 0, 100 and 200, plus targeted related-finding search. Historical source newly added; no substantive update or post-last-run development is claimed.

Publication history

  1. 2026-09-12Local Government · Issue 073 resources
Read preserved resource versions (JSON)

Stable resource ID: atlanta-311-resident-experience-study-2025