Operator Insights

Run it beside another tool1

1. 

   James ReggioCTOBrex

   Organization-wide. Fintech company. Internal AI spotlights, described January 17, 2026.

   Latent Space, published Jan 17, 2026Listen to the episode (opens Latent Space in a new tab)

   Observation

   At Brex, operations, finance and people teams show off agents they built in Glean and ChatGPT.

   Takeaway

   Give non-engineering teams a regular slot to show the agents they build; Brex's adoption runs through those sessions.

   What happened

   Reggio reports that Brex's internal AI spotlights are mostly people in operations, finance and the people team showing how they build agents in ChatGPT or on Glean, or a new use case they found, alongside a chat stack each employee picks.

   “For the most part, it’s folks in VTMs, ops, finance, the people organization showing off how they’re building agents in ChatGPT or on Glean or how they just found some new use case that they thought was helpful.”

   • Podcast episode (Latent Space)
   • No usage figures

   Deployment details

   Organization-wide. Fintech company. Internal AI spotlights, described January 17, 2026.

   • Podcast episode (Latent Space)
   • No usage figures

   Team

   Operations, finance and people teams

   Workload

   Agents for internal workflows

   Product version

   Glean agents beside ChatGPT

Adopted it3

1. 

   Becky WilsonLeads employee digital experienceExpedia Group

   Organization-wide. Travel company with several AI entry points. Described at Glean's conference, published October 14, 2026.

   Glean blog, published Oct 14, 2026Read the write-up (opens Glean blog in a new tab)

   Observation

   Expedia pitched Glean to a team for summarizing quarterly business reviews, and the team built a queryable replacement for the deck instead.

   Takeaway

   Watch what teams build before you prescribe use cases; at Expedia, one team turned a summarization pitch into its own replacement for the QBR.

   What happened

   Expedia routes broad questions that need company context toward Glean while keeping other AI tools. Expedia pitched Glean to one team as a way to summarize its quarterly business reviews and pull action items from the transcript, and that team instead built a full replacement for the QBR that people query.

   “One of the clearest lessons from Expedia's rollout: watch what people actually build, rather than prescribing it for them. They described trying to pitch Glean to a team as a way to summarize their quarterly business reviews and pull action items from the transcript. That team ended up building a complete replacement for the QBR itself, in which all the information that normally would have gone into a deck now lives in Glean. Users can query that pool of information directly before the meeting, which is then spent discussing open questions rather than sitting through a presentation.”

   • Vendor customer story
   • Described on Glean's blog

   Deployment details

   Organization-wide. Travel company with several AI entry points. Described at Glean's conference, published October 14, 2026.

   • Vendor customer story
   • Described on Glean's blog

   Workload

   Quarterly business reviews and company-context questions

   Product version

   Glean

1. MotiveJun 4, 2026Motive engineering blogAt Motive, a Glean agent writes executive account briefs from Slack, Docs, and Snowflake.

   

   Jay RosenthalDirector of Business IntelligenceMotive

   Business intelligence team. Account briefs for sales and customer success executives.

   Motive engineering blog, published Jun 4, 2026Read the write-up (opens Motive engineering blog in a new tab)

   Observation

   At Motive, a Glean agent writes executive account briefs from Slack, Docs, and Snowflake.

   Takeaway

   Let Glean handle synthesis and access control for documents and chat. Rebuild access control for warehouse data an agent can reach.

   What happened

   Rosenthal and BI Platform Manager Shashi Gupta built an agent-to-agent setup. A Glean agent runs federated search over Slack and Google Docs, calls a Snowflake Cortex agent for account metrics and call transcripts, then writes one brief. Glean enforced each source's access rules out of the box, while the team built row-level access by org hierarchy and account ownership for warehouse data. They keep analysis instructions in Glean and have Snowflake return raw data, because passing prompts between agents lost context.

   • Public engineering write-up
   • Architecture described
   • No metrics reported

   Deployment details

   Business intelligence team. Account briefs for sales and customer success executives.

   • Public engineering write-up
   • Architecture described
   • No metrics reported

   Team

   Motive business intelligence team

   Workload

   Account history and open-issue briefs before renewals and meetings

   Product version

   Glean Agents with federated search, calling a Snowflake Cortex Agent

2. Andy ChenMar 10, 2026Andy Chen on SubstackAt Abnormal, Glean found the right documents, and judgment calls still needed a separate context layer built on its search API.

   

   Andy ChenBuilt the go-to-market question botAbnormal

   Go-to-market team. Abnormal had recently started using Glean.

   Andy Chen on Substack, published Mar 10, 2026Read the write-up (opens Andy Chen on Substack in a new tab)

   Observation

   At Abnormal, Glean found the right documents, and judgment calls still needed a separate context layer built on its search API.

   Takeaway

   Use Glean for retrieval, and write down the routing and judgment rules separately before an assistant answers customer-facing questions.

   What happened

   Chen was asked to build a bot that answers sales reps' customer questions, such as whether a feature ships next quarter or how long data is kept after churn. Chen calls Glean excellent at finding documents and used it heavily while onboarding. Retrieval alone would confidently answer questions that should go to the security team, and reps had already given customers wrong answers on them. Chen ran 20 parallel agents with Glean's search API for about two days to write a cited context repository. It held 6,000 commits, 1,020 files, and 11 domains, including routing rules such as sending data-retention questions to security.

   “For reference, we recently started using Glean…”

   • Public practitioner write-up
   • Repository size reported

   Deployment details

   Go-to-market team. Abnormal had recently started using Glean.

   • Public practitioner write-up
   • Repository size reported

   Team

   Abnormal go-to-market reps

   Duration

   About two days of agent runs, six months into the project

   Workload

   Answers for sales reps on roadmap, competitors, and data retention

   Product version

   Glean search API called by 20 parallel agents

   Measurement

   Commits, files, and domains in the context repository; the write-up gives no accuracy score for the answers

Reported by others1

1. 

   Steve JanuarioCIOBill.com

   Organization-wide. Fintech company. Phased rollouts and pilots, described in an August 28, 2025 webcast recap.

   Opsin, published Aug 28, 2025Read the report (opens Opsin in a new tab)

   Observation

   Bill.com's Glean, Gemini and Copilot rollouts surfaced misconfigured access rights and sensitive onboarding files.

   Takeaway

   Audit who can open what before Glean indexes your systems; Bill.com found its gaps through the rollout.

   What happened

   Bill.com adopted Glean early alongside Gemini and Copilot, in phased rollouts and pilots with safeguards. Januario describes how the AI search tools unearthed misconfigured access rights, open Jira boards and onboarding documents with sensitive data.

   • Second-hand news report (Opsin blog, from a webcast with Bill.com's CIO)
   • Recap by an AI data-security vendor

   Deployment details

   Organization-wide. Fintech company. Phased rollouts and pilots, described in an August 28, 2025 webcast recap.

   • Second-hand news report (Opsin blog, from a webcast with Bill.com's CIO)
   • Recap by an AI data-security vendor

   Workload

   Enterprise search across company systems

   Product version

   Glean beside Gemini and Microsoft Copilot