Brands have spent years asking whether consumers trust AI. The third episode of the 2026 Media Outlook series flips that question entirely: Does AI trust your brand?

That reframing, introduced by a senior commerce leader at one of the world's largest consumer goods companies, sets the tone for a conversation that goes well beyond the usual retail media talking points. 

The discussion features Jenna Levin, Senior Director of Global Digital Commerce at Colgate-Palmolive, who oversees the company's e-commerce capabilities and Amazon go-to-market strategy across global markets. Nagarajan Chakravarthy, Chief Digital Officer at iOPEX Technologies, whose mandate spans agentic AI enablement and helping brands build advertising as an operational discipline. Danilo Tauro, Co-Founder of CartographAI, hosts the conversation.

Fragmentation Is Not a Strategy Problem

The conversation opens with a tension every brand-side operator recognizes: investing across six or seven retail media networks, each with its own attribution logic, measurement methodology, and definition of success.

Naga's framing cuts through the noise immediately. "Don't misdiagnose fragmentation as a strategic problem," he said. "It's a fundamental infrastructure problem." Each network has its own attribution secret sauce. Treating that as something to be solved at the planning level, rather than the data layer, is where most brands waste the most time.

His prescription is direct: build your own measurement layer, own your KPIs, and stop outsourcing the interpretation of your performance to the networks that generate it. The shift is from media buying to media operating.

Stop Searching for One Metric. Start Aligning on One Outcome.

Jenna extended the measurement argument in a direction that will resonate with anyone managing retail media across a complex internal stakeholder structure. The search for a single number that satisfies every team, every market, and every budget owner is a dead end.

"Fundamentally, it is about accepting the multimetric reality," she said. Different teams will track different inputs. What cannot differ is the North Star output metric, whether that is market share growth, sales lift, or a media ROI figure that the whole organization agrees to stand behind.

What makes this actionable rather than theoretical is the technology now becoming available to support it. Jenna pointed to AI and clean rooms as the infrastructure that will soon enable brands to test correlations between input and output metrics in real time, rather than waiting for an annual marketing mix model to confirm what the team suspected 12 months earlier.

The Question Nobody Is Asking: Does AI Trust You?

The sharpest insight in the episode comes when Jenna introduces a concept that reorients how brands should think about the next phase of commerce.

"We talk a lot about consumers trusting AI. But I also think we need to talk about AI trusting brands."

The example is concrete. If Amazon's Rufus or Walmart's Sparky cannot verify a brand's product data, the agent will not recommend it. It is not a question of consumer preference or media spend; it is a matter of data structure. An AI agent operating without confidence in your product information will route around you entirely.

"We need to begin thinking about treating AI as one of our most valuable customers."

That means designing PDPs for both human shoppers and AI agents, structuring product data to be API-accessible with rich attributes. Jenna also identifies the second shift: from keywords to context. Consumers are no longer searching in two or three words, they're using eight, building full-sentence queries like "best whitening toothpaste and build me a routine for sensitive teeth." Brands that embed this context into their content are building discoverability infrastructure that compounds. The ones that wait are building a gap no media budget will close later.

The Case for an Agentic In-House Agency

Naga took the argument into execution territory by framing it directly at the problem's operational scale. A single brand can have a thousand product pages spread across seven retail media networks. Keeping the content discoverable and optimized for both human shoppers and AI agents simultaneously is not a human-scale task.

"Why would you not use AI for AI?" he asked.

The concept of an agentic in-house agency follows from that logic. An agent that monitors performance across all seven channels, flags SKUs exhibiting anomalies, generates reallocation scenarios, and surfaces those recommendations to a brand or media operator streamlines the decision-to-action cycle in a way no traditional workflow can match. The human is still in the loop, it’s just that the loop itself moves faster.

Data Is the Oil. What You Build on Top of It Determines Everything Else.

Naga closed with a metaphor worth carrying forward. Data is the oil. Everyone already knows that. What the industry has not yet fully named is what sits on top of it. "AI orchestration is the new soil," he said. It’s the layer that makes it possible to grow from existing data.

The host drew the larger arc together in his summary: Leading brands are no longer optimizing a single channel. They are building commerce operating systems. Those systems connect media, data, content, and organizational structure, and they measure success in terms of business outcomes rather than channel metrics.

That distinction between running campaigns and building infrastructure is the clearest signal of where the gap between leaders and followers is opening up.

Watch Episode 3 of the 2026 Media Outlook for the full conversation.