Benchmark wars are over. Claude Code just proved it — the AI products winning right now aren't the ones with the best models, they're the ones that know their customers best. Context is the new moat.
Juan Sequeda has spent 20 years solving the problem most product teams don't even know they have: your AI is only as powerful as your business's ability to predict what a customer wants to do and why. That intelligence isn't in the model — it's buried in your data. And the secret to unlocking it isn't writing better skills files or crafting smarter prompts. It's re-architecting how your business knowledge is structured, connected, and made available to AI. Juan shows you exactly how.
Product teams who've made this move are seeing accuracy improvements of over 50%, and every new use case they ship compounds on the last. In this episode of the Product Impact Podcast, Juan introduces his three-layer knowledge framework — business metadata, technical metadata, and the mapping layer that connects them — and shows how this foundation transforms what your AI can deliver. You'll leave with a clear starting point, a way to tie your AI investment directly to business outcomes, and a mental model for how the best product teams are pulling ahead. This will only grow as we depend on agents and governance becomes more critical.
In this episode you'll learn:
➡️ Why context — not model quality — is now the primary driver of AI product performance
➡️ The three-layer knowledge framework that gives AI a shared language across your entire organization
➡️ Three concrete first steps to build your context foundation starting tomorrow
➡️ How to tie every AI initiative directly to your company's top OKRs and earn lasting executive buy-in
➡️ Why knowledge-first teams compound their advantage — each new use case gets faster and more powerful
Thank you for listening to the Product Impact Podcast (formerly Design of AI) — Prove impact. Improve impact. Scale impact.
Go to productimpactpod.com to rate the impact of AI products you use at work.
Hosted by:
Arpy Dragffy Guerrero — https://www.linkedin.com/in/adragffy/
Brittany Hobbs — https://www.linkedin.com/in/brittanyhobbs/
Support the show: subscribe, share this episode with a product leader, and leave a rating/review—it’s how this podcast reaches the teams building what comes next.
Subscribe for frameworks + AI strategy resources: https://productimpactpod.substack.comBrought to you by PH1 (https://ph1.ca) — an AI strategy consultancy specialized in improving the measurable success of AI products.
About our guest
Juan Sequeda is Principal Scientist and Head of the AI Lab at data.world, now part of ServiceNow. He has spent 20 years at the frontier of knowledge graphs, ontologies, and semantic architecture — focused on one question: how do you give AI a genuine understanding of your business so it can deliver answers you can actually trust?
His lab's research proved that pairing knowledge graphs with LLMs improves enterprise question-answering accuracy by over 50% — findings that helped define the industry's thinking on context and AI reliability. He co-founded Capsenta (acquired by data.world), coined the concept of "context wars," and recently published his landmark LinkedIn series: "20 Lessons from 20 Years of Building Ontologies and Knowledge Graphs."
He also co-hosts Catalog & Cocktails, one of the most respected podcasts in the data community, and publishes regularly on LinkedIn and Substack.
Resources
➡️ Juan Sequeda on LinkedIn: https://www.linkedin.com/in/juansequeda
➡️ Catalog & Cocktails Podcast: https://data.world/podcasts/catalog-and-cocktails
➡️ Juan's Substack: https://juansequeda.substack.com
➡️ "20 Lessons from 20 Years of Building Ontologies and Knowledge Graphs" — https://www.linkedin.com/posts/juansequeda_i-finished-posting-my-20-lessons-from-20-activity-7429147437681864704-C7ki/
➡️ Software Wasteland — Dave McComb
➡️ The Data-Centric Revolution — Dave McComb
Netherlands