Skip to main content
raf-winterpacht
Level 3
May 5, 2026

Reflecting on Our Summit AMA Session on Content, Commerce, and the Agentic Web

  • May 5, 2026
  • 3 replies
  • 67 views

It’s hard to believe it’s been TWO weeks since I was with my fellow Adobe Advocates, ​@manavpadhariya and KrishnaKalyan Gorthi. We were sitting together at Adobe Summit and talking about the importance of some major shifts happening right now in the industry.

In case you missed it, here is a quick summary and some valuable takeaways:

After quick introductions, we highlighted that while humans might be willing to tolerate a five-second page load, AI bots simply don’t have that kind of patience. They will skip slow or poorly structured sites, and your brand may as well be invisible. They must be performant. In order to build "computational trust" and prevent hallucinations, brands need to design for Machine Experience (MX) just as much as User Experience (UX). This means leveraging semantic HTML, using context-aware schema markups, and well-optimized commerce catalogs. By use of tools like Adobe Sites Optimizer for performance and LLM Optimizer for AI visibility signals—and serving tailored content at the Edge—organizations can ensure they are properly understood, citable, and recommended by AI models while maintaining an excellent human interface and experience.

Top Takeaways

  • Structure is Everything: AI bots don't have “eyes”; they rely entirely on structure. Even better, content should be self-describing and built with semantic HTML so that AI engines can accurately read and represent your brand without hallucinating.

  • Design for Two Personas: You are no longer just serving human customers with nice animations and fancy transitions. Your content, metadata, and commerce catalogs (incorporating basic, advanced, and semantic layers) must be optimized to answer the intent of both human visitors and agentic traffic.

  • Audit and Govern: The journey to the agentic web starts with a thorough audit of your current platform. Governance and keeping a "human in the loop" remain critical to managing brand identity and ensuring that the content AI generates or references remains accurate and trustworthy.

We’d love to hear about how you’re getting organized with your products and content...what are your biggest challenges? Any tips to share about how to optimize for the Agentic Web? Let us know!

3 replies

tekmera
Level 2
May 5, 2026

Great recap! The "structure is everything" point hit home because I lived a smaller version of it on my own site.

I built my site on Lovable, a no-code React SPA builder. Looked fine to humans. It was pure client-side rendering though, so Google and the AI crawlers just got an empty HTML shell. After months live, only 1 of 15 pages was indexed.

Migrated to Next.js on Vercel for server-side rendering. Every page is real HTML on first crawl now. The whole thing took an afternoon and the indexing problem was permanently fixed.

A bit humbling to find out a nice-looking site was effectively invisible to the systems that increasingly decide whether anyone finds you. I know Lovable has recently fixed this issue, so “machine readable” is definitely top of mind in the industry.

David Kershaw | Workfront & Fusion Architecture | Tekmera | https://tekmera.ai/work-with-us/adobe-practice
Jacinda E
Community Manager
Community Manager
May 5, 2026

Thank you so much for sharing this recap, ​@raf-winterpacht! The concept of ‘computational trust’ is something I haven’t really heard about in these terms before. Just as a random thought, I’m curious about your opinion (as well as ​@manavluhar and ​@gkalyan): Do you see a future where brands will need dedicated MX roles or teams, similar to how UX became its own discipline? Or do you think this is something that would become an embedded responsibility across existing content/engineering/UX/ other functions?

raf-winterpacht
Level 3
May 13, 2026

@Jacinda E I think it is absolutely going to be a need, where a dedicated role will need to be focused on MX. One of the titles I’ve heard as of late include a “Forward Deployment Engineer”, and I see this being a part of the larger integration of AI-first initiatives. Great question!

Raf Winterpacht - Adobe SME, Adobe Professional, Adobe Expert
manav
Adobe Champion
Adobe Champion
May 6, 2026

It was a fantastic experience sitting down with ​@raf-winterpacht and ​@gkalyan  at Summit.

@Jacinda E, that is a brilliant question. I actually see MX following the exact same trajectory as UX. Right now, it is mostly an embedded responsibility. We are asking our current architects, SEO experts, and content strategists to "think about the bots" But as the agentic web matures, the complexity will absolutely demand dedicated MX Architects and Strategists. The structural demands of building computational trust are just too specific to treat as a side-task for long.

This ties perfectly into the stakes for commerce right now. Building on Raf's point about designing for two personas, we are seeing a 50+% decrease in brand organic traffic as users begin their journeys inside AI engines rather than traditional search.

When managing massive B2B/B2C catalogs with thousands of SKUs, basic product data isn't enough for the agentic web. You have to focus on the semantic layer, capturing product intent, contextual relationships, and true omnichannel consistency. For anyone wondering where to start, standardizing your data using the UCP standard for commerce catalogs is a great first step before layering on tools like LLM Optimizer or building out a Brand Concierge.

To your point about future roles, this is exactly where a dedicated MX focus will become necessary. I’d love to hear how enterprise brands are managing the governance of their catalog attributes today to ensure both human and agentic traffic get the right experience!

#MagentoMan