Scattered training across Learning Management System, SharePoint, and videos; enterprise search surfaced current content

Overview

A manufacturing firm’s learning content was scattered across the Learning Management System (LMS), SharePoint libraries, and recorded lunch?and?learn sessions. Employees struggled to find current materials, repeated outdated training, and flooded Learning & Development (L&D) with requests for sessions that already existed. Intelligex deployed enterprise search across the LMS, SharePoint, and recorded sessions, normalized metadata, and introduced curator?approved recommendations by role and site. People discovered current, relevant content in one place, managers guided teams without one?off requests, and redundant session demand declined—while the LMS, SharePoint, and recording tools remained in place.

Client Profile

• Industry: Manufacturing (production, maintenance, quality, and corporate functions)
• Company size (range): Multi?plant footprint with centralized L&D and site?based training coordinators
• Stage: LMS for compliance and skills training; SharePoint for work instructions, SOPs, and job aids; recorded sessions in Microsoft Stream/meeting platforms; ad hoc search and duplicated content
• Department owner: Human Resources & People Ops (Learning & Development and HRIS)
• Other stakeholders: Operations and Maintenance, EHS/Safety, Quality, IT/Collaboration Platforms, Security & Privacy, Plant Managers, Union or Works Council where applicable, Internal Communications

The Challenge

Operators and technicians needed to find the right content quickly. Instead, LMS search returned incomplete results, SharePoint pages surfaced older PDFs alongside revised SOPs, and recorded sessions lived in separate channels with titles that didn’t match course names. People bookmarked whatever they could find, and teams repeated training that was already available, sometimes using superseded materials.

Requests piled up. Managers asked L&D for refresher sessions that already existed as recordings or micro?courses. Site trainers rebuilt materials because they couldn’t tell which version was current. Compliance modules were re?taken simply because people couldn’t locate completion records or updated equivalents. L&D triaged email threads to point people to the right assets, which left little time for content improvements.

Consistency and trust suffered. Each plant organized SharePoint differently, titles and tags varied by author, and recorded sessions were rarely transcribed or tagged. Employees weren’t sure which source to trust, and supervisors couldn’t see whether a recommended module was the latest. This created risk for safety and quality training where only current content should be used.

Why It Was Happening

Content and metadata were fragmented. The LMS tracked courses and completions, SharePoint held SOPs and job aids, and recorded sessions lived in separate channels. None of those systems shared a common taxonomy, and each had its own search limited to that platform. Without normalized metadata and a single search that respected permissions, people guessed based on file names and recency rather than on curated relevance.

Governance was light. Versions were published without retiring older files, tags were inconsistent across plants, and recorded sessions lacked transcripts or summaries. Recommendation lists lived in spreadsheets or email. Without a curator workflow and clear ownership for deprecations and tags, the path of least resistance was to rebuild or re?request training rather than to discover what already existed.

The Solution

Intelligex delivered an enterprise search and curation layer that indexed learning content across the LMS, SharePoint, and recorded sessions, normalized titles and tags, and presented curator?approved recommendations by role and site. The service reconciled duplicates, flagged stale items, and boosted current, canonical sources. Role?based views mapped employees to content relevant to their job family, line, or plant. Recording transcripts and summaries became searchable, and a light curator workflow allowed L&D and site trainers to approve recommendations and retire outdated materials. The design used Microsoft Graph search to index SharePoint and OneDrive content (Microsoft Graph Search), SharePoint APIs for metadata updates (SharePoint REST API), and portable standards like xAPI for training signal interoperability (xAPI). Recorded session handling aligned to platform capabilities such as Microsoft Stream.

• Integrations: LMS catalog and course status; SharePoint libraries for SOPs, job aids, and specs; recorded sessions and transcripts; HRIS for role, location, and manager; identity provider for security trimming; collaboration tools for notifications; audit logs to the compliance repository.
• Metadata normalization: Canonical titles and summaries; version/effective date fields; tags for job family, equipment, process, and plant; synonym lists for common terms; duplicate detection and merge suggestions.
• Recommendations service: Role? and site?aware content cards; curator approval queue; freshness and quality scores that boosted canonical, current content; “related items” for cross?training.
• Search and permissions: Security?trimmed results that matched source access; filtering by role, plant, topic, and format; clear badges for “current,” “under review,” and “retired.”
• Content hygiene and governance: Stale content alerts to owners; deprecation workflow with redirects; transcript generation and indexing; accessibility checks aligned to WCAG.
• Dashboards and reporting: Top searches and zero?result queries; recommendation engagement; duplicates and stale items by plant; content owner workload and turnaround.
• Privacy and security: Role?based authoring and curation; minimal PII in analytics; encryption in transit and at rest; immutable logs for indexing, edits, and approvals.

Implementation

• Discovery: Inventoried LMS catalogs, SharePoint libraries, and recording channels; sampled duplicate and stale items; mapped job families, equipment types, and plant?specific needs; reviewed identity and permission models; gathered Safety, Quality, and Privacy requirements for versioning and access.
• Design: Authored the metadata schema (title, summary, effective/retire dates, tags, owner); defined synonym lists and duplicate rules; designed the role? and site?aware recommendation model; specified curator workflows for approvals and deprecations; planned search UI, dashboards, and evidence exports; set accessibility and transcript policies.
• Build: Connected LMS, SharePoint, and recorded session sources; implemented indexing and security trimming; built normalization, duplicate detection, and freshness scoring; created curator queues and approval screens; generated transcripts and summaries where supported; instrumented dashboards and logging.
• Testing/QA: Ran in shadow mode to compare enterprise results to platform searches; validated permissions and security trimming; exercised duplicate merges and redirects; piloted recommendations in two plants with L&D curators; tuned synonym lists, tags, and freshness thresholds based on real queries.
• Rollout: Launched read?only search and role?based recommendations first; enabled curation and deprecation workflows plant by plant; kept LMS and SharePoint publishing flows unchanged; added redirects from retired files and updated site navigation; tightened boosting rules after curator confidence grew.
• Training/hand?off: Delivered quick guides for employees on finding content and using filters; trained curators and site trainers on approvals, deprecations, and tags; briefed content owners on stale alerts and transcript checks; updated SOPs for publishing and retirement; transferred schema, synonym lists, and dashboards to L&D and HRIS under change control.
• Human?in?the?loop review: Established monthly content councils with L&D, Safety, and Quality to review top queries, zero?result terms, duplicates, and retired items; recorded decisions with rationale and effective dates; fed updates back into tags, synonyms, and governance rules.

Results

Employees found the right content without guessing where it lived. Search returned current courses, job aids, and recordings with clear labels and role?appropriate filters. Recommendations by role and plant surfaced what mattered most, and duplicates pointed to the canonical source with redirects. Managers sent fewer ad hoc requests because they could point teams to a single, trusted view.

L&D shifted effort from hunting links to improving content. Curators approved recommendations, retired obsolete files, and saw where queries failed. Recorded sessions became searchable with transcripts, so refresher topics didn’t require new facilitation. Compliance modules and safety content were easier to locate, which reduced repeated enrollments and off?catalog requests. Core systems remained; the new layer connected them with unified search, governance, and role?aware recommendations.

What Changed for the Team

• Before: People searched LMS, SharePoint, and recording channels separately. After: One search returned security?trimmed, current results across all sources.
• Before: Outdated PDFs and recordings stayed in circulation. After: Curator workflows retired old items and redirected to canonical content.
• Before: Managers requested repeat sessions. After: Role?based recommendations and searchable transcripts satisfied most needs.
• Before: Titles and tags varied by author and plant. After: A shared schema, synonyms, and tags made content easier to find.
• Before: L&D answered the same “where is it?” questions. After: Dashboards showed gaps and guided curation instead of inbox triage.
• Before: Accessibility and versions were inconsistent. After: Transcripts, WCAG checks, and effective/retire dates became part of the flow.

Key Takeaways

• Index where content already lives; unify LMS, SharePoint, and recordings with one permissions?aware search.
• Normalize metadata; shared titles, tags, and effective dates beat file?name guessing and local conventions.
• Curate recommendations; let role? and site?aware suggestions guide learners while owners approve and retire.
• Make freshness visible; label current vs retired, and redirect rather than leaving duplicates to linger.
• Treat transcripts and accessibility as content; searchable text and WCAG checks improve discovery and inclusion.
• Integrate, don’t replace; keep existing tools—add search, governance, and recommendations across them.

FAQ

What tools did this integrate with? The solution indexed SharePoint and OneDrive content via Microsoft Graph Search and updated metadata through the SharePoint REST API, pulled catalogs and status from the LMS, and surfaced recorded sessions and transcripts using platform capabilities such as Microsoft Stream. Training signals remained portable through standards like xAPI. Identity governed access and role?based recommendations.

How did you handle quality control and governance? A shared metadata schema, synonym lists, and deprecation rules lived under change control with L&D and HRIS as owners. Curator queues approved recommendations and retirements. Stale alerts prompted owners to review content, and redirects pointed old links to canonical items. Every index, approval, edit, and deprecation wrote to an immutable log with timestamps and actors.

How did you roll this out without disruption? Enterprise search launched read?only alongside existing platform searches. Curator workflows and retirements started with a few plants while LMS and SharePoint publishing stayed the same. As confidence grew, boosting rules and redirects tightened, and site navigation linked to the new search. No rip?and?replace of LMS or SharePoint was required.

How were roles and sites used for recommendations? The service used HRIS attributes—job family, location, equipment, and line assignment—to filter and boost content by relevance. Role?based cards and plant tags highlighted what each audience needed first, with the full catalog still searchable. Sensitive or plant?specific content followed source permissions.

How did you handle outdated content and duplicates? Duplicate detection flagged near?matches across sources. Curators merged or redirected to a canonical item and retired others with a visible “retired” badge. Effective/retire dates in metadata and redirects kept older links from resurfacing without context.

What about accessibility and recorded sessions? Recording transcripts and summaries were generated where supported and indexed for search. Content checks encouraged clear headings, alt text, and link clarity aligned to WCAG. Learners could skim transcripts to confirm relevance before launching long videos.

Did this change compliance tracking? Compliance assignments and completions remained in the LMS. Enterprise search and recommendations helped learners reach the right items faster, and xAPI signals or LMS logs continued to serve as the system of record for completion.