Koddi this week released a 54-page industry report finding that visibility inside AI-generated shortlists is rapidly becoming the central investment priority in commerce media, displacing traditional placement and impression-based advertising as the primary competitive arena for brands and commerce media networks alike.

The report, titled The State of Agentic Commerce (Media): How Agentic AI is Reshaping Control, Value and Monetization, was announced on June 10, 2026. It draws on fieldwork conducted by The Human Instinct, an insight and strategy consultancy based in London, in March and April 2026. The research surveyed 750 consumers - 250 each in the US, UK, and Germany - plus 150 senior commerce media professionals split evenly across 25 supply-side and 25 demand-side respondents per market. Industry participants included commerce media networks, retail media networks, retailers, brands, advertisers, media agencies, and ad tech platforms.

What the headline numbers show

The single most striking finding for the marketing community is that 84% of commerce media leaders said they would invest in opportunities specifically designed to increase brand visibility inside AI-generated answers and recommendations. That figure signals a meaningful reorientation of budget intent away from click-based and impression-based formats toward what Koddi frames as a new layer of paid influence.

Adoption of AI agent infrastructure among industry professionals is already substantial. According to Koddi, 70% of commerce media leaders are currently using AI agents to analyse data, recommend actions, and automate campaign execution. Yet the same group expresses clear caution about removing humans from consequential decisions. Only 3% of respondents said they favoured environments in which humans are primarily limited to oversight and compliance. The dominant preference, held by 60% of industry respondents, is for a "shared: AI proposes, humans approve most big decisions" model. A further 20% prefer a mostly human approach with AI handling only analysis and suggestions.

These numbers matter because they reveal an industry that has moved well past early-stage testing without embracing full autonomy. According to Koddi, most respondents characterize their organizations as currently operating in what the report calls Wave 2 - agentic shopping and decisioning - and expect to shift toward Wave 3, defined as orchestration and system-level monetization, within the next 12 to 18 months.

Nicholas Ward, president and co-founder of Koddi, put the structural argument plainly: "Commerce media is entering a new phase where influencing decisions inside AI-driven shopping environments is becoming increasingly valuable. The industry is moving beyond traditional media buying models toward systems focused on decisioning and visibility inside AI-mediated shopping experiences. AI is becoming more integrated into shopping and discovery, meaning the companies that can maintain visibility and measurable performance inside the systems shaping consumer decisions will have the competitive advantage."

Three waves, and where the market sits now

Koddi's analytical framework divides the adoption of agentic AI in commerce into three distinct phases. Wave 1 covers AI embedded within existing shopping and advertising environments - personalization engines, automated bidding, optimization within established ad formats. This phase improves performance within current structures without altering how decisions are made or how value is captured.

Wave 2 involves AI actively participating in decision-making. Agents assist with search, comparisons, and filtering; recommend products, brands, and services; and support both consumers and marketers with decision support and purchasing on their behalf. According to the research, most organizations in all three markets sit in Wave 2 today.

Wave 3 is defined by AI orchestrating decisions across systems - coordinating experiences across multiple platforms and networks, integrating data and inventory and decision logic, and enabling monetization based on access, influence, and outcomes. The report projects the market will move toward Wave 3 within 12 to 18 months. At that point, the key competitive question shifts from "Who has the inventory?" to "Who is included in the decision process?"

The transition, however, is not uniform. According to the research, nearly half of US respondents said AI is integrated and scaled across multiple functions in their organizations, compared with a quarter in Germany and one in five in the UK. Among industry professionals, 70% of those in the US said they are "very familiar and already using" AI agents, against 48% in the UK and 32% in Germany.

Consumer attitudes: assistance yes, autonomy no

The consumer findings add a layer of structural tension that matters directly for advertisers. Three-quarters of US consumers (75%) say they are comfortable with AI helping them choose what to buy. That number drops to 68% in the UK and 59% in Germany. But comfort collapses sharply when the question shifts to autonomous action: only 20% of all consumers are comfortable allowing AI to independently choose and act on their behalf.

These are not fringe statistics. Across every product and service category the research examined - retail, groceries, travel, utilities, banking, investments, health, tech, auto - no more than one in seven respondents said they were happy to let AI choose and act without intervention. The majority prefer AI that suggests options while humans decide, or AI that provides basic information without making choices.

The boundary, according to Koddi's analysis, is "a graduated boundary based on risk, reversibility, and intimacy." Consumers are most relaxed about AI involvement in retail, travel, groceries, and tech - categories that feel more familiar and reversible. They are least relaxed about banking, investments, and health, where the perceived consequences of errors are high.

Specific use cases illuminate this pattern. Sixty-four percent of consumers say they favour AI suggesting brands they might not usually buy. Fifty-five percent are comfortable with AI reordering everyday items when running low. Almost 50% want AI to plan a vacation. But automatic repeat purchasing without asking each time is acceptable to only 8% of respondents, the lowest-scoring behaviour tested.

The trust architecture consumers want is procedural rather than emotional. The top three safeguards they demand, rated essential by more than four in five consumers in each market, are secure data handling, human checking for high-value decisions, and the ability to undo, cancel, review, and change choices. Making a purchase without explicit approval was named as the single biggest trust reducer by 55% of respondents. Using personal data in a way that feels unclear was second at 39%. Recommending products because a brand paid for placement came third at 35%.

The monetization question

One of the more nuanced findings in the report concerns paid commercial influence inside AI recommendation systems. Consumers are not absolutists on this issue. More than half - 56% - say it is acceptable, or acceptable if clearly labelled, for brands to pay for shortlist inclusion, first position, or default fulfilment. The critical condition is transparency: hidden commercial influence is rejected far more strongly than visible and labelled commercial influence.

The German market presents a paradox that the report examines in some detail. German consumers are the least comfortable overall with AI taking autonomous action in commerce, yet they display comparatively higher tolerance for paid influence when it is clearly disclosed. According to Koddi's analysis, nearly three in ten German consumers say paid influence is acceptable if clearly labelled, versus around one in five in the US and UK. The report interprets this as a different underlying trust model - not more enthusiasm for AI, but greater acceptance of overt commercial rules than of invisible optimization.

For the advertising and commerce media industry, this distinction matters because it implies that emerging monetization formats built around AI shortlist inclusion can be commercially viable - provided they are disclosed. The research does not establish whether existing regulatory frameworks in each market would require such disclosure, nor does it examine how disclosure should be implemented technically.

Industry investment intentions and pricing models

The industry data on investment intentions provides some of the most operationally specific findings in the report. More than three-quarters of advertisers described themselves as "very likely" or "somewhat likely" to invest across all six categories of agentic commerce products the researchers tested: data and feeds integration fees, paid AI-visibility or ranking products, sponsored shortlist or sponsored recommendation slots, default fulfilment partnerships, cross-agent orchestration tools, and agent-specific measurement and diagnostics.

On projected budgets, roughly a third of advertisers expect to allocate between $50,000 and $249,000 in the next 12 months to agentic commerce products, with an equal share expecting to allocate $250,000 to $999,000. A further one in five planned investment of $1 million to $5 million. The report notes this level of commitment requires internal justification, governance, and expected return - signalling the category has crossed what it calls "the credibility threshold."

The most popular pricing models among respondents are cost per sale or order (chosen by 53%), fee for guaranteed eligibility and access (35%), and subscription or platform fee and cost per click (both at 31%). Traditional CPM, at 22%, underperformed relative to outcome-linked and access-linked models. This preference pattern supports the report's broader argument that monetization is shifting from impressions to influence and outcomes.

When asked which monetizable unit will be most important in an agentic AI environment, 37% of industry respondents chose "access and eligibility in the agent's knowledge and decision set," followed by "completed outcomes such as sales, leads, and bookings" at 29%. These responses sit in tension with traditional media buying assumptions that prioritize audience reach and frequency.

The source of funding is also notable. Sixty-one percent of advertisers said they would draw from performance and paid search budgets to fund agentic AI in commerce media - the largest single category. The UK came in highest at 66%, the US at 62%, and Germany at 54%. This has direct implications for paid search budgets and for how agentic AI is already reshaping the ad buying layer.

Measurement as the central constraint

On both the consumer and industry sides, measurement and attribution are identified as critical structural barriers. According to Koddi, 92% of respondents plan to invest in agent-specific measurement and diagnostics, and 80% plan to invest in measurement tools tied to AI agent-mediated shopping journeys. These are not aspirational statements. The report frames measurement readiness as a gating condition for scaling investment.

Thirty-three percent of industry respondents named "measurement and attribution of AI agent-mediated journeys" as a major blocker of investment in agentic AI for commerce media - equal to legal and compliance frameworks for AI usage and behind only internal skills and governance for AI decisioning at 38%. The performance bar before scaling is also clearly set: most respondents require a 10% to 25% improvement in return on ad spend or cost per acquisition before expanding an agentic offering beyond a test scenario.

Eric Brackmann, VP of Commerce Media at Koddi, described the consumer-industry dynamic in the research: "Consumers are increasingly embracing AI as a shopping companion, but they're not ready to hand over complete control. Our research points to a future where AI plays a larger role in product discovery, evaluation and decision-making. However, transparency and human oversight still remain essential to building trust. As AI-generated recommendations become a more influential part of the shopping journey, visibility within those experiences is emerging as a critical new frontier for commerce media. Commerce media networks have an opportunity to develop new monetization models and advertiser offerings designed specifically for agentic environments. The ones that succeed will be the ones that balance innovation, performance and consumer trust as AI reshapes how purchase decisions are made."

How CMNs and advertisers diverge

The 150-person industry sample was divided equally between commerce media network and retailer respondents on the supply side, and brands, advertisers, media agencies, and ad tech platforms on the demand side. Despite broad alignment on the direction of travel, the report identifies four areas of divergence that will shape how value is distributed.

On where value sits, CMNs are more likely (36% versus 23% of advertisers) to see monetization concentrating in access to data and feeds that make a brand eligible to be considered. Advertisers are more likely (19% versus 5%) to see value in orchestration and optimization layers across networks. The report characterises this as CMNs seeking to control entry into the system, while advertisers seek to control how the system operates.

On what gets funded, CMNs expect brands to invest most heavily in data and feed integration fees (91% "very" or "somewhat likely to invest" versus 81% among advertisers) and sponsored shortlist or recommendation slots (85% versus 72%). Advertisers are almost twice as likely to prioritize lower-funnel conversion and sales, at 45% versus 24% for CMNs. CMNs are more than three times more likely to predict heavy investment in upper-funnel awareness, at 28% versus 8%.

On evidence of impact, CMNs are significantly more likely to identify new-to-brand customer impact as the most important proof point (59% versus 32% of advertisers). Advertisers prioritize longitudinal test effectiveness at 51% versus 39%, and independent third-party verification at 32% versus 21%.

On risk perception, CMNs express higher concern about regulatory and legal risk exposure (47% versus 36%) and brand safety and suitability (43% versus 28%). Advertisers prioritize data privacy and security at 60% versus 51%.

The broader structural conclusion the report draws is that value is concentrating not among those with the most inventory, but among those most deeply integrated into decision systems. As Amazon has also been formalizing its position in agentic commerce, the competition for inclusion in AI decision layers is being waged across the entire ecosystem simultaneously.

Winners, losers, and the decision economy

The report closes with a framework for anticipating which types of organizations will gain from the shift to agentic commerce and which will lose ground. According to Koddi's analysis, winners will include data-rich and performance-driven brands, AI platforms and large language model providers, integrated commerce media platforms that connect data, decisioning, and execution, and measurement and analytics layers that can demonstrate incremental impact and ROI.

The categories identified as likely losers are impression- and click-based media models, surface-dependent media formats that rely on being seen on a page, feed, or app, fragmented point solutions that cannot operate across systems, execution-heavy human-first agency models that fail to evolve, and any brand or platform that is not integrated into the systems that make decisions.

This maps closely to structural dynamics that PPC Land has covered across The Trade Desk's integration with Koddi's platform, the IAB's 9.5% ad growth forecast driven by agentic AI investment, and the broader shift in consumer trust questions around AI shopping agents. The question of who controls the most commercially valuable decisions in an agent-mediated commerce environment draws a fragmented answer from the research: brands (41% of CMN respondents, 40% of advertiser respondents), large consumer AI platforms and LLM providers (41% versus 32%), and retailers and CMNs (35% versus 25%). No single actor holds a decisive lead.

Timeline

Summary

Who: Koddi, a commerce media technology platform based in Fort Worth, Texas, whose partners include Babylist, Booking.com, Cars.com, Fanatics, Gopuff, Grubhub, Kohl's, and Kroger, released the research. The survey was designed and conducted by The Human Instinct, a London-based insight and strategy consultancy, in March and April 2026. Respondents included 750 consumers (250 each in the US, UK, and Germany) and 150 senior commerce media leaders split evenly across supply-side and demand-side professionals in the same three markets.

What: Koddi released a 54-page research report finding that 84% of commerce media leaders intend to invest in AI recommendation visibility products, that 70% are already using AI agents for campaign execution, and that 92% plan to invest in agent-specific measurement and diagnostics. The report also finds that only 20% of consumers are comfortable with AI acting autonomously on their behalf, and that 56% of consumers accept paid brand influence in AI shortlists provided it is clearly disclosed.

When: The research fieldwork was conducted in March and April 2026. The report was released on June 10, 2026.

Where: The research covers the US, UK, and Germany. Industry respondents were sourced equally from all three markets. Consumer respondents were sampled at 250 per market. Koddi is headquartered in Fort Worth, Texas. The Human Instinct is based in London.

Why: The report addresses a structural shift in how brand visibility, campaign measurement, and media monetization function as AI agents increasingly mediate the product discovery and purchasing process. Commerce media networks face pressure to evolve beyond selling placements toward influencing and governing the AI-driven decisioning systems that determine which brands consumers see and ultimately choose.