Connected scattered enterprise knowledge into a pre-launch LLM Assistant that delivered 25% productivity gains

25% improvement in sales & field enablement efficiency • 60%+ adoption in under 3 months 

The Challenge

In the intense pre-launch phase of the AWS European Sovereign Cloud, content and knowledge were being created in silos across dozens of teams: sales pitch decks, product BOMs, enablement training materials, engineering announcements, marketing blogs, regulatory summaries, and more.

A large Slack channel with hundreds of members (sales, field, enablement, SMEs, product managers) tried to bridge the gap, but it quickly became noisy and inefficient. Valuable content existed — but it was fragmented, hard to find, and inconsistent. Teams wasted time searching, waiting for answers, or risking outdated information in customer conversations.

As the central content operations lead building the aws.eu marketing platform, I was already focused on governance, structured content, taxonomy, and search strategy. I saw firsthand how enterprise silos — even with good intent — create friction. No one was connecting the dots. I realized this was a perfect opportunity to turn siloed content into something truly useful.

PROJECT STAGES

1. Identified the productivity pain caused by siloed content and proposed an internal AI assistant using Amazon Q
2. Acted as the “content connector” — curated, audited, and structured a unified knowledge repository from 20+ siloed sources (BOMs, pitch decks, blogs, announcements, training videos, website pages, regulatory docs, etc.)
3. Applied content governance and strategy expertise to transform fragmented material into a coherent, high-quality knowledge base
4. Designed quality guardrails, response boundaries, tone/voice, audience alignment, and assistant personality
5. Ran a focused pilot with cross-functional early adopters and executed rapid iterations
6. Added safety measures and launched internally before sovereign cloud GA

PROJECT STAKEHOLDERS

• Sales & Field teams
• Enablement & Campaign teams
• Product Managers & Subject Matter Experts
• Local market SMEs
• Distinguished Engineers

Enterprise environments often produce excellent content in isolation. The missing piece is someone with a content strategist’s lens to connect the dots.

I brought that mindset:

• Audited and consolidated siloed assets into a single, structured repository
• Applied taxonomy and semantic metadata principles (the same ones I used for the aws.eu platform)
• Layered in strong governance guardrails so the assistant would only deliver authoritative, on-brand, compliant answers
• Defined clear rules: what it could answer, when to say “I don’t know,” tone for sales audiences, and strict boundaries on regulatory topics

This turned scattered, tactical content into a unified, intelligent resource.

I started with a targeted pilot of early adopters chosen for expertise diversity. Initial error rate on complex queries was 34–40%. Through several rounds of refinement — strengthening sources, tightening guardrails, and improving prompts — I reduced it to under 5%.

Prominent disclaimers ensured responsible use: “This tool boosts productivity but should not replace human judgment or official documentation.”

• 25% improvement in sales and field enablement efficiency — instant, trustworthy answers replaced slow searches across silos.
• 60%+ adoption rate within the first three months.
• Dramatically reduced Slack noise and repetitive cross-team questions.
• Strong testimonials highlighting how the tool made previously siloed knowledge instantly actionable.
• Established a repeatable model for turning enterprise content silos into AI-powered assets.

All delivered before the public launch, demonstrating proactive value during peak execution pressure.

In large enterprises, great content is often created in silos. What turns it into real business value is a content strategist who can see the bigger picture, connect the dots, govern the quality, and make it accessible at scale. When you combine that mindset with generative AI, you unlock exponential productivity gains while maintaining accuracy and brand safety.

AI/GenAI Content Workflows • Breaking Down Content Silos • Knowledge Base Curation & Structuring • Content Governance & Quality Guardrails • Retrieval-Augmented Generation (RAG) • Pilot Design & Iterative Testing • Cross-Functional Collaboration • Productivity Enablement • Search & AI-Ready Content Strategy