Two days at the Venetian with 14,000 attendees from more than 60 countries, and one key takeaway. The conversation has moved from generative AI to agentic AI, and from experimentation to measurable outcomes.

Jensen Huang put the frame most clearly on stage with Shantanu Narayen. Generative AI produces information. Agentic AI produces work. As Huang said, AI is "finally valuable" because it now executes against goals rather than just generating answers. He used his own company as proof: 100% of NVIDIA's software engineers are now supported by agents, and they are busier than ever, not displaced. Faster experimentation cycles. More ideas tested. The same thing happened to radiologists when AI made scan analysis superhuman. Demand went up.

That distinction supports and accelerates the transformation roadmaps we have seen playout across retail over this past year. Here is what stood out.

1. AI is a demand channel, not just a productivity tool

Adobe's 2025 holiday data showed AI-driven traffic to retail sites grew 693 percent year over year. In a separate Summit panel, Adobe shared that visitors arriving through agentic search convert 4.4 times higher than other channels.

This belongs on every CMO and CIO scorecard. If your content is not structured, syndicated, and optimized for LLMs, you are losing visibility at the moment of intent. Brand discoverability has become a balance sheet issue, not a marketing experiment.

DICK'S Sporting Goods made this concrete on the Day 2 keynote. CMO Emily Silver and CTO Vlad Rak walked through their partnership with Adobe on Brand Concierge, turning the digital storefront into a coaching tool that understands athlete intent and guides them from discovery to decision. The stack pairs Adobe Experience Platform, Real-Time CDP, Journey Optimizer, AEM, Workfront, and Firefly Custom Models. The point is not the products. The point is that personalization at scale only works when the underlying data and content foundation can feed it.

2. Campaigns are out. Journeys are in.

Fanatics gave the most useful operating lesson of the week. Gayatri Moray described how the company moved from three disconnected systems and channel-based teams to a unified profile and lifecycle-based model on Adobe Journey Optimizer. The number that mattered: 50 to 60 percent of implementation time was spent on architecture and data modeling before a single journey was built.

That ratio explains why most personalization programs stall. AI is a multiplier on a clean foundation. Without the foundation, propensity models and recommendations have nothing reliable to act on. With it, every interaction compounds. Purchase suppression, real-time cart validation, channel deduplication, and adaptive follow-ups happen by design.

This is the structural shift retail boards should be funding. Not more campaigns. Not more tools. A different operating model.

3. The content supply chain is now infrastructure

Walmart's Kate Christiansen made the case that the content supply chain is the prerequisite for creative excellence at scale. Single front door for intake through Workfront. Brand-aware AI tagging in AEM. Workload balancing across teams. Firefly generating derivative variations against a master AI asset, with smart reprompting, so a single approved creative produces infinite on-brand outputs.

Best Buy's Stacy Aase and Katie Phillips made a similar argument with their activation-first DAM blueprint. The goal is to stop treating asset management as a storage locker and start using it as the engine that drives creative speed, governance, and personalized delivery.

The Adobe and NVIDIA partnership on 3D digital twins points to where this goes next. Pixel-perfect virtual product replicas enable infinite composite imagery, lifestyle scenes, and personalized variants without re-shooting. For retailers carrying tens of thousands of SKUs across markets, that is a margin lever, not a creative novelty.

4. Use cases beat vendor pitches

Ulta Beauty had the most practical playbook of the week. CMO Kelly Mahoney and CTO Mike Maresca walked through how they moved the enterprise from AI experimentation to customer experience impact, anchored in a Who, What, Where, Why, How framework that ranks use cases by business impact and feasibility. The framework is unglamorous, and that is exactly why it works. It forces priority, owners, and measurable outcomes before budget is spent.

CVS Health applied the same discipline in a regulated environment. Liz Kaufman and Calisha Anderson showed how an integrated stack of Adobe Target, Journey Optimizer, and Workfront delivers AI-driven, HIPAA-compliant Rx personalization at scale. Same lesson, different vertical: integrated foundation, governed AI, defined use cases, measurable impact.

Amazon Ads, with LeapPoint and Omnicom, told a parallel story on the marketing operations side. Unified data governance and intelligent automation are how you reclaim speed and eliminate the manual chaos that personalization demands keep creating.

5. The new competitive divide

A few patterns ran through every customer story.

Retailers with unified profiles, real-time data pipelines, and clear operating models are scaling outcomes. Those without them are stuck in pilots. AI compresses multi-step manual workflows down to minutes, but the gain is not doing the same work faster. The gain is rethinking processes end-to-end and redeploying talent from reporting to strategy and optimization. Brand integrity and governance become harder, not easier, as AI-generated content multiplies. And the executives winning this transition are pairing CMO and CTO as a single decision unit. DICK'S and Ulta both put their CMO and CTO on stage together. That tells you something about how this gets bought and built.

What this means for the next 12 months

Three questions every retail leader should be asking right now.

Is my content infrastructure structured and syndicated for LLM discovery, or am I still optimizing for a search era that is already shrinking?

Have I funded the data and operating model work that AI needs to compound on, or am I trying to bolt agents onto fragmented systems and hope for the best?

Are my CMO and CTO operating as one team with shared accountability for customer experience, or are they still owning separate slices?

The retailers who answer these clearly will own the next cycle. The ones who keep treating AI as a tooling decision will spend the year explaining why the pilot did not scale.

Would love to hear your key takeaways and compare notes on what this looks like inside a Canadian enterprise context.

Sessions worth your time on demand:

Originally published on LinkedIn as “Notes from Adobe Summit 2026: Retail’s New Operating Model”, April 30, 2026. Reproduced here as written; formatting simplified. Third-party statistics are as cited in the original and are not independently verified on this site. Read the original on LinkedIn ↗