Case 03 · 2026
From Fragmented Standards to an AI-Ready Knowledge Hub
Making enterprise knowledge easier to find, understand and reuse.
- AI assistant
- By dimensions
- By region
- Drawings
- Legacy view
At a glance
- Context
- Internal knowledge environment with ~30–40 technical standards pages
- Timeframe
- Jan – Sep 2026
- Role
- UX Project Lead
- UX Champion
- Product-adjacent responsibilities
- Status
- Live since mid-September 2026
The challenge
The problem was not a lack of information.
Knowledge was spread across numerous standards pages created over time, with inconsistent naming, structure and information logic. Finding an answer often meant moving between pages, documents and other sources — or asking someone who already knew where to look.
The aim was not to create new technical standards. It was to organize existing knowledge around the questions users brought to it, with clearer context, reusable content structures and room for future AI-assisted access.
Find the right standard quickly
Long manual searches across pages, PDFs and other sources.
Understand regional applicability
“Can I use this standard in my region?” was often not answerable directly.
Navigate inconsistent structures
Different naming and page structures across the library.
Move between sources
Users had to check several systems or ask colleagues.
Know what is current
Valid or obsolete status was not always clear.
My role
I prepared an internal user study, synthesized the findings and refined them through follow-up interviews. As UX Project Lead, I led the target information architecture and navigation concept, defined reusable page structures, and coordinated UX, content and technical workstreams. My product-adjacent responsibilities included clarifying scope, prioritizing needs, aligning the target direction and discussing dependencies — not formal Product Owner responsibility or ownership of the wider platform.
UX
- Prepared the survey and refined findings through interviews
- Synthesized and prioritized needs
- Led sitemap and navigation concept
- Defined a standardized page structure
Product-adjacent
- Scope protection and requirement clarification
- Target state and roadmap alignment
- Dependency and feasibility discussions
Alignment
- Stakeholder alignment
- Preparing user validation
- Defining the AI assistant use case from the user perspective
Provided expertise
- User Research
- Research Synthesis
- Information Architecture
- Findability
- Product Thinking
- Scope Definition
- Prioritization
- Knowledge Management
- Stakeholder Alignment
- AI Enablement
Glimpse of the process
A research-first process — the structure had to come before the interface and before AI.
- 01
Discover
Internal user survey on search behaviour and information gaps, followed by interviews to check the findings against daily work.
- 02
Define
Problem framing and scope: what belongs to the project, and what doesn’t.
- 03
Structure
Navigation based on user questions rather than the inherited page tree, with a familiar route retained during transition.
- 04
Design
A reusable page template and a consistent content model.
- 05
Build
Review and reorganize content, identify redundancy and align naming during migration.
- 06
Validate
Stakeholder and user validation.
- 07
Iterate
Refining with data-quality and technical dependencies in mind.
Information architecture · Conceptual shift
From knowing where to look to knowing what you need
Page-led navigation
- Different page names
- Uneven content structures
- Scattered context
The user has to interpret the structure before finding an answer.
Need-led navigation
- Find relevant guidance
- Check applicability
- Access supporting resources
Clear entry points lead into a consistent knowledge structure.
A bridge, not a forced switch
Keep a familiar route available while users move towards the new navigation.
Key insights
Research themes
Insights were translated into underlying user needs — not into a feature list.
Regional validity
A key need was to understand where guidance applies without searching other sources or relying on personal contacts.
Findability
Users wanted to start from practical requirements and context rather than having to learn the existing page hierarchy.
Consistent architecture
The same naming, hierarchy and metadata logic on every page.
Content clarity
Clearer valid/obsolete status and coverage — partly a governance question.
AI-assisted access
Research showed openness to guided, dialogue-based support as another route to the same information — not a replacement for clear navigation.
Content model · Shared foundation
One consistent foundation, more than one way in
- Purpose
- What this guidance helps with
- Context
- Where it applies and what to check
- Status
- What is current and who can clarify it
- Related guidance
- Where to continue
- Browse
- Recognizable categories and consistent pages
- Ask
- A guided, AI-assisted route to the same knowledge
Both routes depend on trustworthy content and clear ownership. Better navigation cannot correct the underlying data.
Conceptual design principles — not an internal template or confirmation of live AI functionality.
Key decisions
“Scope protection: not every user problem was a UX problem.”
Some needs belonged to master data, governance, technical standards or other system owners. UX could make them visible, but did not claim to solve them.
Structure first, AI on top
- AI-assisted access
- Search & entry points
- Consistent content model
- Information architecture
“Transition design: keep the old path open.”
Alongside new strategic entry points, the legacy structure stayed accessible — so existing users were not forced into a completely new mental model overnight.
“AI cannot compensate for poor information architecture.”
The AI assistant was designed as a second way to reach the same information — and intentionally depended on the new structure and content model being in place first.
Constraints & trade-offs
This was not a greenfield redesign. The solution had to balance:
User needs
Prioritized from research.
Platform limitations
The existing enterprise wiki and publishing layer restricted layouts, interaction patterns and direct AI integration.
Data quality
Regional availability and smart search are only as good as the underlying data.
Technical feasibility
AI integration and structure options had to be validated.
Organizational ownership
Clarifying data validity depended on other owners. Unresolved dependencies had to stay visible rather than being treated as design decisions.
Phased coverage
Content volume and uneven data readiness required prioritizing an initial usable scope rather than assuming complete coverage everywhere.
Implementation status
Post-launch usage data is not available, so this is reported as status rather than measured impact.
Launched
- — Scalable information architecture: a more consistent, maintainable structure across 30–40 standards pages
- — Navigation redesign: clearer categorization and task-oriented entry points, with the familiar structure retained for transition
- — AI-ready foundation: consistent content organization and entry points prepare the hub for AI-assisted knowledge access
- — Cross-functional alignment: customer needs, content requirements and technical constraints balanced in one future direction for the platform
- — Go-live: the new knowledge hub launched in mid-September 2026
Dependencies & follow-up
- — Post-launch adoption and usage measurement are not yet reported
- — Complete regional coverage is not confirmed; it depends on verified underlying data
- — AI access was designed as an additional route; full integration and performance are not confirmed by the supplied evidence
- — Master-data quality and validity remain responsibilities of the relevant data owners
What this case demonstrates
Enterprise UX is often less about designing an ideal interface and more about balancing user needs, information quality, organizational ownership and technical constraints.