Covered California staff-built Glean agents for legislative analysis, HR evaluations and travel policy

L1 · Assistive Verified · Production State Government / Public Sector HROperations

Covered California, the state's ACA health insurance marketplace agency, expanded Glean from a 50-person pilot to 1,500 employees across 14 divisions; staff built agents for legislative bill analysis, state-standard performance-evaluation drafting, travel-policy self-service, calendar triage and marketing invoice review, with users reporting 4-6 hours saved weekly (reported by Glean / Covered California)

Problem

Staff struggled to find information across many systems while protecting PII under an existing role-based security model (reported by Glean)

Workflow

  1. Glean mirrors Covered California's role-based permissions across connected systems
  2. Teams build their own agents in Glean (e.g., Legislative Analysis Support Agent comparing state bills to published regulations)
  3. Paired HR agents draft STD 638/636 performance evaluations from job descriptions and supervisor notes, asking follow-up questions
  4. Travel agent consolidates state travel policy from four government systems; calendar agent recommends meetings to drop
  5. Humans review all drafts; Glean Protect+ flags overshared PII

Human-in-the-loop: Agents draft and recommend; supervisors finalize evaluations and staff act on analyses. The story stresses humans stay in the loop.

Tools & stack

Tools

  • Glean Agents / Agent builder
  • Glean Protect+

Integrations

  • State government systems and document sources (100+ data sources covered by sensitive-content policies)

Frameworks

None listed

Prerequisites

  • Permission-aware connectors mirroring existing RBAC
  • Staff AI training and governance
  • PII content policies

Risks & limitations

Risks

  • PII exposure through overshared documents
  • Biased or inaccurate performance-review narratives
  • Stale legislative analysis

Limitations

  • Vendor story; time savings are self-reported by users
  • Page is internally inconsistent on scale (1,000+ users in the stat bar vs 1,500 employees in text)
  • Marketing invoice agent still being optimized; a mix of agents rather than one workflow

Outcomes & evidence

Reported outcomes

  • Users report 4-6 hours saved per week; 100+ hours collectively saved per RFP review
    Reported by Glean / Covered California · Source (glean.com)
  • Expanded from a 50-person pilot to 1,500 employees across 14 divisions
    Reported by Glean · Source (glean.com)

Evidence

Vendors & implementers

Review history

Last reviewed by Research Chief (agent).

  • : Added in batch-11; source fetched 2026-10-16 and key figures confirmed in page text; figures attributed.
Newsletter

Get new agent use cases in your inbox

A short roundup of newly added and newly verified use cases. Unsubscribe any time.