Library Resources Employees Actually Find
Cross-library search with source-status checks so dead links don't waste time, use-case-grounded recommendations for new hires and topic-driven needs, and a governance dashboard that surfaces what's broken, what's stale, and what employees are searching for that doesn't exist yet. Five admin-only write tools, all confirmation-gated.
Why Library Resources Don't Reach Employees
Libraries Agent attacks the four specific failures that turn a curated library into a dead archive — without making admins re-curate from scratch.
Employees Search The Wrong Library Or Give Up
The expense form is in Finance Library, but the new hire searches HR Library, doesn't find it, and emails their manager. Each library has its own scope; nobody knows the mental map. Cross-library search has to be the default, not an admin-only feature.
Half The Links Are Dead And Nobody Notices Until Someone Tries
A linked tool changes URLs, a vendor SSO endpoint moves, a deprecated form gets archived. The library link still points to the old place. Without a live source-status check on every item, every result is a coin flip.
"Where Do I Start?" Has No Recommended Path
A new hire opens HR Library and sees 47 items. Where to begin? Without a use-case-grounded recommendation ("I'm a new hire — what's first?"), every employee builds the same mental ranking from scratch, badly.
Admins Don't Know What's Missing Until Employees Stop Asking
"Parental leave checklist" gets searched 28 times this month — zero results, zero conversions to anything else. Without a gap report, the admin can't see the unmet demand; employees just go around the library.
Curated Resources Go Stale And Nobody Re-Reviews Them
An admin curated the Onboarding Library two years ago — handpicked links, training videos, policy summaries. Half the content references a tool the company stopped using, a benefit plan that was replaced, or a process owner who left. Nothing has changed in the library because there's no signal pushing the admin back to review it.
The Same Question Gets Asked In Five Apps Instead Of One Library
"Do we have a template for X?" lands in #ask-hr, then #ask-legal, then a DM to the ops manager, then a forum post. Each owner answers individually because nobody knows the same question already came in three times this week. A library that surfaced the top unanswered queries would let admins promote the right resource — once.
Libraries Agent At A Glance
Libraries AI
Cross-library search · use-case recommendations · governance + gap reports.
Inside Libraries Agent — The Actual Capabilities
Every block below maps to a real tool the agent uses against your Libraries data. Eight read tools surface items, libraries, recommendations, and governance health. Five write tools are in RISKY_TOOLS and require admin role plus explicit confirmation before they touch a record.
Cross-Library Search With Source-Status Checks
Employees ask in plain English; the agent searches across every library they have access to. Results include item type (simple_link, file, form, survey, wiki, post, image, video), library, category, and a source-status flag so dead links surface as broken before the employee clicks through.
- search_library_items — natural-language search across all accessible libraries, with optional library and item-type filters.
- list_libraries — every library the user can see, with item counts and descriptions.
- browse_library — drill into a specific library to see its categories and items.
- get_item_details — full item details including the source-status flag (whether the linked resource is still active).
Use-Case Recommendations And Library Statistics
"I'm a new hire — what should I read first about benefits?" gets a ranked list of resources grounded in the use case. The agent reads library structure, recency, and item popularity to propose a starting path. For admins, get_library_stats exposes counts, type breakdowns, and overdue-review items.
- recommend_resource — natural-language use case in, ranked list of most-relevant items out.
- get_library_stats — counts, item type breakdown, recently updated items, libraries overdue for governance review (admin dashboard).
- Permission-aware — only items the user can already access surface in recommendations.
- Freshness-aware — last-updated and source-status both feed into the ranking, so dead and stale items rank lower.
Governance Health, Knowledge Gaps, And Admin Writes — Gated
get_governance_report rolls up open findings, top-affected libraries, broken links, and stale content. get_gap_report surfaces zero-result searches over the last N days — the content employees are looking for that doesn't exist yet. Admin writes (create, update, delete, move, create category) are gated by RISKY_TOOLS and admin role.
- get_governance_report — open findings by severity, top affected libraries, broken-link rate, overdue reviews (admin only).
- get_gap_report — zero-result searches over the last N days (default 30) with top unmet queries.
- 5 RISKY admin writes — create_library_item, update_library_item, delete_library_item, move_library_item, create_category. All require admin role + explicit confirmation.
- Audit trail on every action — read or write, every call logs the requesting user, the tool, and the affected record.
Outcomes Teams Can Measure
The agent's job is to make every library item findable across the whole catalog, surface dead and stale content before employees hit it, and tell the admin what's missing — not just what's broken. Measure against your pre-agent baseline.
- Cross-library search hit rate — share of queries returning at least one item the user clicked through to.
- Broken-link surface rate — share of dead links surfaced by source-status before the employee clicks (vs after).
- Knowledge-gap closure — top zero-result queries converted to created items within 30 days.
- Recommendation adoption — share of new-hire / topic-driven sessions where employees act on the agent's ranked list.
- Governance-finding burndown — open findings closed per week, by library and severity.
Five Admin Writes, Always With A Human In The Loop
Libraries Agent's RISKY_TOOLS list is 5 entries — every write is admin-role-gated AND confirmation-required. Employees use the agent in read-only mode; admins get the same agent with write authority, and every create / update / delete / move surfaces for explicit approval before it touches a record.
- 5 RISKY admin writes — create_library_item, update_library_item, delete_library_item, move_library_item, create_category. All gated.
- Admin role required — non-admin users cannot invoke any write tool, even with confirmation; the agent rejects at the permission layer.
- Permission-aware reads — non-admin users only see items their role allows in the underlying library; cross-library search respects access scopes.
- Delete is destructive — delete_library_item is explicitly flagged as unable to be undone, and the confirmation prompt surfaces that warning.
WHAT TEAMS TRY INSTEAD
The four alternatives — and why none of them see all twelve libraries, your roles, or your retention
Most ops and knowledge teams reach for one of these four before considering an embedded agent. None of them stick because none of them have native access to SOPs, Policies, Training, Forms, KBs, Marketing KB, and the other content libraries the platform already runs.
ChatGPT, Claude, or Copilot pasting content in
General-purpose AI on copied snippets
- Searches across all twelve libraries at once instead of one paste at a time
- Honors per-library role permissions — non-admins never see content their role can't access
- Curates, archives, and reassigns ownership — not just answers questions
Glean, SharePoint Knowledge Copilot, Notion AI search
Vendor-trapped enterprise search inside another silo
- Already has every MangoApps library indexed — no separate connector setup, no nightly recrawl
- Knows what type each result is (SOP vs Policy vs Training) and routes the next action accordingly
- Performs library admin actions — archive stale items, reassign owners, flag duplicates — not just retrieve
A custom RAG search on a separate vector DB
Engineering's build that drifts the moment content updates
- Reads live content, not yesterday's embedding snapshot — no re-index lag after a Policy edit
- Inherits per-library ACLs from the canonical source — no parallel access model to keep in sync
- Ships as a tool surface inside the chat people already use, not a separate search portal nobody opens
The manual fallback — "ask the SOP team"
A ticket, a Slack DM, or a hunt across twelve libraries
- Deflects routine "where's the latest version of X" lookups the curation team shouldn't be a router for
- Surfaces duplicate, stale, and ownerless content so the team can clean up before audit
- Cross-library search means employees stop opening the wrong app to look for the right document
PLATFORM ADVANTAGE
Libraries Agent inherits everything the platform already runs
An enterprise-search vendor has to plumb each of these. The agent gets them for free because the platform already does.
Twelve native libraries
Reads SOP Hub, Policy Hub, Training, Forms, KB, Marketing KB, Procedures, Templates, and the rest as first-class sources — no connectors to wire, no permissions to mirror.
Per-library role enforcement
Honors the role scoping each library already enforces — the same ACL that gates the UI gates the agent's reads and writes.
Audit trail & retention
Every read and write lands in AiApiLog with the same retention and eDiscovery posture as the rest of the platform.
Confirmation gates on writes
Archive, reassign-owner, and delete tools all require explicit confirmation — destructive actions cannot fire from a bare ask.
Live content, no re-index lag
Queries the canonical library directly — no nightly embedding refresh, no stale snapshot, no drift between the agent and the policy of record.
RubyLLM-grounded model tiering
Nano / small / medium / standard tier selection routes routine lookups to cheap models and reserves the big ones for cross-library reasoning — automatically, per call.
INDUSTRY FIT
Industries where library sprawl moves the most weight
Libraries Agent shines wherever multiple content surfaces hide the same answer.
Healthcare
Pulls the latest clinical SOP, policy version, and required training certification in one answer — and flags the duplicate procedures hiding in two different libraries.
Manufacturing
Surfaces stale safety SOPs and orphaned procedures before audit so plant supervisors aren't training on last year's version.
Retail
Cross-references promo policy, store SOP, and the right training module in one ask so associates aren't toggling between four apps mid-shift.
Financial Services
Tags every retrieved policy with version, owner, and last-reviewed date so compliance can prove the answer cited the current rule of record.
Hospitality
Routes the right brand-standard SOP and training module to property managers who need an answer in the next five minutes, not the next five days.
Public Sector
Runs entirely inside FedRAMP-eligible deployment options with full audit logging — no AI traffic leaving the tenant boundary.
WHY MANGOAPPS WINS
An embedded library agent beats a chatbot, an enterprise-search add-on, or a custom build on every axis
The argument finance, IT, compliance, and ops all share — and the one a horizontal AI or single-vendor assistant structurally cannot answer.
Cheaper than the alternatives
No per-seat Glean license, no Copilot add-on, no six-month custom RAG build, no extra knowledge-ops headcount to curate twelve libraries.
More secure
Every read and write logs to AiApiLog with the same retention as the rest of the platform. No second vendor with a full embedding of your policy and SOP corpus.
Easier to deploy
Already deployed if any library is enabled. Turn the agent on, point it at the role scopes you already configured, and it's running the same day.
Easier to use
Lives inside the chat and library UI — no separate search portal, no context-switching, no remembering which app holds which document.
Easier to manage
Per-business confirmation gates, write-tool toggles, and admin-only enforcement sit in the same admin console as every other app's settings.
Easier to extend
Shares the agentic tool framework with every other MangoApps agent. New library types, new admin actions, and new curation signals ship as tools, not rewrites.
AI is actually better
A horizontal or vendor-trapped AI can return a search snippet. Only Libraries Agent can also reassign owners, archive duplicates, and enforce per-library role permissions — and act on all of it.
At our head office within HR, marketing, and finance, we use ChatGPT quite a bit, and MangoApps AI Helper is a useful tool for us. As I go around to the stores, I’ve been showing them how the AI Intranet Helper can find things and answer plain language questions about our resources, and they’re quite excited about that.
Alison van Zijl
Head of Human Resources
Full House
Customer Success
Related Customer Stories
Customer Case Studies
Customer Case Studies
Video Case Studies
Customer Case Studies
Customer Case Studies
Customer Case Studies
Frequently Asked Questions About Libraries Agent
13 tools — 8 reads and 5 admin writes. Reads include search_library_items, list_libraries, browse_library, get_item_details, recommend_resource, get_library_stats, get_governance_report, get_gap_report. Writes (admin only, all gated) are create_library_item, update_library_item, delete_library_item, move_library_item, create_category.
All 5 admin writes are in RISKY_TOOLS and require explicit user confirmation. delete_library_item is destructive and the confirmation prompt surfaces that warning. Non-admins can't invoke any write — the permission check fails before the model ever sees the tool as available.
search_library_items only returns items the user already has access to in the underlying library. There's no privilege-escalation by going through the agent — if a library is hidden from the user in the app, it's hidden in agent results too.
get_gap_report tracks zero-result searches over a configurable window (default 30 days). The top unmet queries surface the content employees are looking for that the library doesn't have — a forward-looking signal for curation, not a backward-looking governance report.
Cross-library search hit rate, broken-link surface rate (surfaced before vs after click), knowledge-gap closure (top zero-result queries converted to items within 30 days), recommendation adoption on new-hire flows, and governance-finding burndown per week.
Let's Talk
Since 2008, we've been building the employee platform for the frontline, earning the trust of 2 million+ users and an NPS of 78.
Why Choose Us?
- AI-Ready Platform: One intelligent place for every employee and workflow.
- Top Security: HITRUST, ISO & SOC 2 certified.
- Exceptional UX: Delightful on mobile and desktop.
- Proven Results: 98% customer retention rate.
Trusted by Legendary Companies:
Prefer to explore first? Ask AI about Libraries AI Agent →