The Challenge
With the company facing the perfect storm of content chaos, and traditional metadata approaches that were not sufficient for its AI ambitions; they needed semantic-level specificity to disambiguate similar terms with different meanings so a Large Language Model (LLM) wouldn’t assemble an answer from the wrong context.
At the same time, they were standardizing authentication and access management with Okta for single sign on (SSO) and needed granular permissions to ensure that the right people saw the right content every time. On top of all of that, the team needed clear content governance, publication workflows, ownership, and quality controls to maintain content health at scale.
The Solution
They turned to Shelf to deploy both pillars of the Shelf Platform:
Pillar 1: Shelf Knowledge Management System (KMS) for robust governance frameworks, content management, and semantic tagging.
Pillar 2: Shelf Content Intelligence for comprehensive diagnostics, automated remediation recommendations, and GenAI enablement.
The solution centered on four critical foundations:
- Diagnostics and content insights: Automatically flags content ROT for quicker identification and elimination.
- Publication workflows: Complete draft-to-archive lifecycle with defined review, approval, and scheduled archival processes.
- Ownership and accountability: Clear content owners with established Service Level Agreements (SLAs) and responsibility matrices.
- Semantic metadata: Moving beyond traditional tags to word- and concept-level context calibrated specifically for GenAI accuracy.
- Access control: Role-based visibility layered over Okta SSO integration, ensuring least-privilege access to sensitive information.
The Journey
Phase 1: Diagnostics-First Discovery
Conducted a comprehensive Content Intelligence content scan within the Shelf Platform, providing ROT scoring and precise de-duplication recommendations across all repositories.
Phase 2: Structured Cleanup Sprint
Targeted removal and consolidation efforts reduced the content library from 25,000+ documents to 9,500 high-quality assets.
Phase 3: Semantic Enhancement
Advanced tagging moved beyond traditional metadata to capture the nuanced context AI systems need to distinguish between related but distinct concepts.
Phase 4: Governance Activation
End-to-end publication workflows and content ownership structures were established, creating sustainable quality controls.
Phase 5: Access Integration
Seamless Okta SSO integration with Shelf’s granular permissions system ensured the right content reached the right people every time.
Phase 6: Continuous Optimization
Search analytics identified gaps, improved findability, and dramatically reduced abandoned searches through iterative refinement.
Why This Matters
They didn’t just “turn on” an AI assistant and hope for the best. They took the right order of operations:
- Establish content health and governance
- Enrich metadata from traditional to semantic, so AI interprets concepts correctly
- Secure access with SSO and permissions
- Then introduce GenAI with confidence
Human Support Accelerated the Change
But it wasn’t only the platform making an impact. Shelf’s Customer Success and Enablement team worked side-by-side with the company to guide the change, accelerate cleanup, and upskill internal owners. One team member was able to eliminate over 60% of ROT in a matter of weeks using Content Intelligence insights and Shelf workflows—evidence of how quickly teams can move when they have the right tooling and coaching.
“I wanted to express our genuine appreciation and gratitude for the Shelf Customer Success Team. They have been nothing short of absolutely amazing to work with in the past six months or so. We were astounded by their deep knowledge of AI. On top of that knowledge, they are easy to get along with and are absolutely professional. We leave every meeting with them happy and know that our voices have been heard as well. Not only have they been able to answer our questions, they have also been able to provide alternative solutions on how we can improve our processes, research, or maintenance. They have been nothing short of invaluable to us and their advice has helped us clear up our content to make a better Shelf experience for our organization. Our relationship with them is one of the reasons why we look forward to our continued partnership.”
The Company’s Corporate Operations Trainer shared their appreciation for Shelf’s Customer Success Team.
The Results
The transformation delivered measurable impact across every key metric:
- 62% of content ROT eliminated (duplicates, obsolete, and low-value documents removed).
- Document library streamlined from 25,000+ to 9,500.
- Significant improvement in successful searches and dramatic reduction in abandoned searches.
- Role-based content delivery implemented via Okta SSO with granular permissions.
- AI-ready content foundation established with semantic tagging, governance, and safety controls for GenAI consumption.
- 2,500+ users now accessing clean, reliable, contextually relevant information.
The Outcome & Future
By modernizing content governance and metadata, the company minimized the risk of AI hallucinations, protected sensitive information, and ensured their teams could trust the answers they received. The result is cleaner, faster, safer knowledge delivery—positioning the company at the forefront of practical, enterprise-ready AI.
This disciplined approach ensures they’re not rushing into AI deployments that need to be rolled back—they’re set up to deliver high-quality answers from the start and sustain that quality over time.
Why Shelf?
Shelf.io is the leading AI-driven knowledge automation platform, helping employees and customer service agents instantly find answers across any channel. Founded in 2017, Shelf powers enterprises like DSW, HelloFresh, Gerber, Glovo, and John Deere, and is ranked #1 in Knowledge Management by Gartner.
See how Shelf can diagnose your content health, eliminate inefficiencies, and prepare your organization for confident AI deployment. Get in touch →
