Glean’s Insights Chat and Admin Chat help organisations run AI deployments with less guesswork. Insights Chat lets admins explore adoption and usage trends in natural language—by team, feature, and time period. Admin Chat provides conversational help for configuration and governance, making it easier to manage policies and settings without specialist tooling.

AI deployments often fail for boring reasons.

Not because the model is weak—but because nobody can answer simple operational questions quickly:

• Are teams actually using the assistant?

• Which departments are stuck after onboarding?

• What changed before usage dropped?

• Which settings matter for safety, and who owns them?

Admin teams end up juggling dashboards, spreadsheets, support tickets, and fragmented “gut feel” updates. That’s exactly the gap Glean’s latest admin experiences are designed to close.

Insights Chat helps you explore adoption and usage patterns simply by asking questions. Admin Chat helps you manage and optimise your deployment in the same way—through natural language, with guardrails.

This guide explains what’s new, how it works in practice, and how to roll it out so it improves governance and adoption (instead of becoming another admin surface area).

Internal link placement: If you’re exploring Glean for enterprise search and Work AI, see Explore Glean → /glean/.

Why AI management needs to change

Most organisations have moved past the “pilot excitement” stage. Now the hard part is operational: scaling usage, proving value, and keeping control.

Traditional administration struggles because:

1. Dashboards answer what happened—not what to do next. You can see a trend line, but it’s still unclear why it moved.

2. Configuration is often locked behind specialist knowledge. Even small changes can require deep product familiarity.

3. Governance teams need defensible reporting. Leaders want KPIs and risk signals they can trust, not anecdotal success stories.

Natural language admin experiences help by removing the translation layer. Instead of turning every question into a dashboard investigation, you can ask what you need in plain English—then move straight to action.

Internal link placement: For a board-level view of governance, link to AI governance for boards → /blog/ai-governance-evolving-board-strategies.

What’s new: Insights Chat vs Admin Chat

Insights Chat: visibility in plain language

Insights Chat is designed for the questions admins ask every week:

• “Which teams are using Glean the most this month?”

• “Where has adoption dipped since last quarter?”

• “Which agents are being used—and by whom?”

• “What feedback themes are coming through?”

Instead of manually filtering dashboards, you use conversational prompts to explore adoption across core product areas.

Admin Chat: operational control without the faff

Admin Chat is built for deployment work: configuration help, guidance on settings, and smoother day‑to‑day administration.

The big shift is not that admin tasks become “non-technical”—it’s that more of the work becomes accessible and auditable. You can reduce admin bottlenecks, standardise how changes are made, and lower the risk of misconfiguration.

What you can measure (and why it matters)

The fastest way to make AI “stick” is to measure what adoption really looks like—then respond with targeted enablement.

Here are the most useful categories to track with Insights Chat, and how to turn each one into action.

1) Adoption and engagement

What to look for: active usage trends, department breakdowns, new vs returning users, and patterns over time.

Why it matters: broad adoption is a leading indicator of value, but it’s also an early warning system for enablement issues.

What to do next:

• If usage is concentrated in a few teams, identify champions and run peer-led sessions.

• If usage drops after onboarding, improve role-based prompts and “first wins”.

Internal link placement: For the people side of AI, link to AI adoption change management → /blog/ai-adoption-change-management.

2) Feature utilisation (what people actually use)

What to look for: which parts of the product are being used most (search, assistant, summarisation, agents, etc.).

Why it matters: many deployments fail because the organisation assumes people are using features they’ve never discovered.

What to do next: run short, targeted enablement focused on the top three workflows each department needs.

3) Agents and automation

What to look for: which agents are used, how widely, and where they drive repeatable outcomes.

Why it matters: agent adoption is where ROI often accelerates—but it can also introduce governance questions.

What to do next: publish an approved agent catalogue, define owners, and create a lightweight review loop.

4) Quality and feedback signals

What to look for: feedback themes, pain points, and patterns that suggest people don’t trust the outputs.

Why it matters: trust drives repeat usage. If people don’t believe the answers, adoption stalls even when the tool is available.

What to do next: prioritise connector coverage, permissions health, and content hygiene—then show users how to ask better questions.

Practical examples: prompts admins can use today

Below are prompt patterns you can copy into Insights Chat and Admin Chat. (Adapt wording to match what your tenant supports.)

Insights Chat prompt examples

• “Show adoption trends for the last 90 days by department.”

• “Who are our power users in Sales and Customer Support?”

• “Which teams have the lowest engagement since onboarding?”

• “What are the most common queries in the last month?”

• “Which agents are used most often, and by which departments?”

• “Summarise feedback themes for the Assistant this quarter.”

Admin Chat prompt examples

• “Help me check the recommended settings for a secure rollout.”

• “What’s the best way to enable a new connector without breaking permissions?”

• “Explain the difference between [setting A] and [setting B] in simple terms.”

• “What should I review before enabling this feature for the whole organisation?”

• “Draft a short admin change note for stakeholders about enabling [feature].”

These prompts do two things well:

1. They reduce admin time spent switching tools.

2. They make administration more consistent—because the questions become repeatable.

How to roll out Insights Chat and Admin Chat safely

A strong rollout isn’t about “turning it on”. It’s about integrating the tools into governance rhythms and team habits.

Step 1: Define your north-star questions

Before you touch settings, agree the questions leadership expects you to answer.

Examples:

• Adoption: “Are we growing usage in priority departments?”

• Impact: “Which workflows are saving time?”

• Safety: “Are we seeing risky patterns or misconfiguration signals?”

Turn these into a monthly reporting template so insights stay consistent.

Step 2: Establish ownership and guardrails

Decide who owns:

• platform configuration

• connectors and permissions health

• adoption enablement

• risk and compliance reporting

Admin Chat reduces friction, but you still need accountability. Put change approvals and comms in a lightweight process.

Internal link placement: For scaling governed AI, link to Enterprise AI at scale → /blog/enterprise-ai-governance-security.

Step 3: Run a 2–4 week “learning loop”

Pick 2–3 departments, define one measurable workflow each, and use Insights Chat weekly to track progress.

Focus on:

• prompt quality

• connector coverage

• enablement content

• stakeholder comms

Internal link placement: Link to Work AI learning loop → /sp/blog/work-ai-learning-loop-nextdoo.

Step 4: Train admins on prompt patterns (not prompt perfection)

You don’t need long prompt workshops. You need a small library of repeatable admin questions.

Create a one-pager:

• top 10 Insights Chat prompts

• top 10 Admin Chat prompts

• what to do when results look off

Step 5: Turn insights into action

Insights are only valuable if they change behaviour.

Build a simple cadence:

• Weekly: adoption + quality check

• Monthly: department review + enablement actions

• Quarterly: governance review + roadmap decisions

Common pitfalls (and how to avoid them)

“We measured adoption, but nothing changed.”

If you track metrics without attaching actions, you’ve built reporting theatre.

Fix: define a small set of triggers. Example: if engagement drops in a priority department, run a targeted enablement session within 10 working days.

“Admins are still the bottleneck.”

If only one person knows how to use Admin Chat well, you’ve recreated the problem.

Fix: document standard questions and a change checklist, then share it across the admin group.

“We improved usage, but trust is still low.”

Trust issues often come from connector gaps, permissions mismatches, or messy content sources.

Fix: prioritise content hygiene and connector coverage, then publish a simple “how to ask” guide tailored to each department.

Where Generation Digital helps

Tools like Insights Chat and Admin Chat are most powerful when they sit inside a well-run operating model.

Generation Digital helps organisations:

• implement Glean with security and permissions aligned from day one

• define adoption KPIs that leadership can trust

• build governance rhythms that don’t slow the business down

• train teams to use AI safely, consistently, and confidently

Summary

Glean’s Insights Chat and Admin Chat bring a practical shift to AI administration: visibility and control via natural language.

If you’re responsible for AI operations, this is the difference between “we think it’s working” and “we can prove it—and improve it.”

Next steps

1. If you’re exploring Glean, start here: Explore Glean → /glean/

2. If you need a governed rollout plan, read: Enterprise AI at scale → /blog/enterprise-ai-governance-security

3. To discuss implementation, adoption, and governance for your environment: Contact Generation Digital.

FAQs

Question: What is Glean Insights Chat?
Answer: Insights Chat is an admin-facing chat experience that lets you explore adoption and usage trends in natural language, helping you understand how Glean is used across teams, features, and time periods.

Question: What is Glean Admin Chat?
Answer: Admin Chat is a conversational admin experience that helps you manage and optimise your Glean deployment—supporting configuration questions, rollout guidance, and day-to-day administration with fewer manual steps.

Question: What should we measure to prove AI adoption is working?
Answer: Track adoption trends by department, repeat usage over time, feature utilisation, and feedback quality signals. The most useful metrics are the ones tied to clear enablement actions.

Question: Do natural-language admin tools replace governance processes?
Answer: No. They reduce friction, but you still need ownership, approvals, and reporting rhythms. The best outcomes come when admin chat is embedded into a clear operating model.

Question: How long does it take to embed these tools into workflows?
Answer: Most organisations see value quickly if they start with a focused pilot (2–4 weeks), standard prompt patterns, and a cadence that turns insights into enablement actions.