Why agentic commerce media will be a parallel universe for human-facing advertising
Opinion
Agentic commerce will not replace traditional marketing. Instead, it will create a complex parallel operating system alongside existing consumer marketing, adding another layer of fragmentation, co-ordination challenges and competing priorities for brands.
Across the industry, agencies are rapidly building AI-driven systems capable of autonomously signalling, planning, buying and optimising media investment.
On the commerce side, retailers such as Amazon and Walmart are launching consumer-facing shopping assistants like Rufus and Sparky, designed to help consumers navigate purchasing decisions with minimal friction.
Much of the current conversation around agentic commerce focuses on efficiency, automation, and replacement. The assumption is that agentic systems will streamline decision-making, reduce human intervention, and fundamentally reshape how brands engage consumers. But the reality is likely to be far more complex.
Agentic commerce will not replace traditional marketing. Instead, it will create a complex parallel operating system alongside existing consumer marketing, adding another layer of fragmentation, co-ordination challenges and competing priorities for brands.
The future of commerce media is not simplification; it’s co-existence.
The dual-target reality
For decades, marketers focused on a single audience: the consumer. Strategies were built around emotional resonance, cultural relevance, memory structures, cognitive bias, and visual distinctiveness. Success depended on understanding how people think, feel, and behave.
Now, however, a second audience is emerging: the machine agent.
Where human-facing marketing is driven by emotion, storytelling, aspiration and trust, machine logic is driven by structured data, API signals, inventory feeds, fulfilment reliability, pricing dynamics and product attributes.
A shopping agent does not care about the emotional nuance of a brand platform, or the nostalgia triggered by a TV campaign. It cares about real-time availability, pricing accuracy, delivery speed, verified product information and predictive confidence.
At the same time, algorithms cannot replicate the irrational emotional loyalty that causes someone to instinctively reach for a particular brand on a supermarket shelf.
This creates a new marketing duality.
Strategies that perform brilliantly in human-facing environments may become almost invisible inside agentic systems. Equally, brands optimised purely for machine recommendation risk becoming functionally efficient but fail to build emotional relevance with people.
Consumers may increasingly rely on AI agents to narrow choices, compare products, and automate routine purchases, but the underlying emotional associations that shape trust, familiarity and brand affinity will still be formed through human experience.
We are already seeing early examples of this shift in consumer-facing shopping assistants, such as Amazon’s Rufus and Walmart’s Sparky, which are designed to simplify increasingly complex purchase decisions. Adoption signals suggest consumers are becoming increasingly comfortable with AI-assisted commerce journeys:
- 54% of consumers are likely to engage with an AI chatbot on a retailer’s website.
- 40% of Amazon’s Black Friday sessions used Rufus.
- Walmart reported that users spent 35% more when using Sparky.
- Generally, chatbots are estimated to increase sales by approximately 67%.
We are also seeing early-stage forms of agentic commerce, such as Amazon’s Smart Reorders, in which platforms anticipate purchasing needs before consumers actively make decisions.
These behaviours are still emerging, but they point towards a future in which machine-assisted purchasing becomes embedded throughout everyday commerce journeys.
Increased fragmentation, not simplification
One of the biggest misconceptions surrounding agentic commerce is the belief that AI will reduce fragmentation across media and commerce ecosystems. In practice, the opposite is more likely.
Alongside existing media channels, brands will need to navigate multiple autonomous ecosystems, each operating with different optimisation logics, data structures and commercial incentives. Rather than consolidating complexity, agentic commerce may intensify it.
Retailer-owned AI assistants such as Amazon’s Rufus, Walmart’s Sparky and future retailer-owned platform-based shopping agents will operate within proprietary environments built on closed datasets and platform-specific algorithms. Brands will not be able to rely on a single optimisation strategy across every ecosystem because each platform will prioritise different signals.
This marks an important evolution for retail media.
Historically, media planning and operational commerce functions were often treated separately. Agentic systems compress these disciplines together because operational signals directly influence recommendation outcomes.
Many brands are not currently structured to co-ordinate media, commerce, technology, supply chain and operational functions at the level agentic systems may require. Yet success in machine-mediated environments will increasingly depend on how effectively those functions work together.
Human behaviour still matters most
Another misconception surrounding agentic commerce is the assumption that automation will diminish the importance of human decision-making. It will not.
Consumers will still seek identity, trust, discovery, and emotional connection from brands. They will still respond to creativity, culture, and experiences that AI systems alone cannot replicate.
In fact, as more purchasing journeys become automated, the moments where consumers actively engage may become even more influential.
The less frequently consumers make conscious purchasing decisions, the more valuable those moments of emotional engagement may become.
Over-optimising for machine conversion risks undermining the long-term brand investment required to influence human preference. Machine agents naturally optimise towards short-term conversion efficiency and measurable outcomes, whereas human brand-building relies on long-term emotional memory structures that cannot always be captured through immediate transactional metrics.
The brands that succeed will not be those that optimise exclusively for agents or exclusively for people. They will be those capable of balancing both simultaneously.
That balance will become increasingly difficult because machine systems reward consistency, predictability, and measurable performance, while human attention often rewards surprise, originality, storytelling and emotion.
These are fundamentally different marketing dynamics.
Agentic commerce therefore looks less like a replacement cycle and more like a coexistence model, where human persuasion and machine optimisation operate in parallel.
The strategic challenge is orchestration
The immediate challenge facing the industry is not simply the adoption of standalone AI tools; it’s orchestration.
Most organisations are not currently designed to manage multiple autonomous systems operating simultaneously across commerce, media, operations, and customer experience. Nor are many prepared for the governance implications of increasingly automated decision-making.
As agentic ecosystems mature, the challenge will not simply be integrating AI into existing workflows. It will be managing the interaction between human strategy, machine optimisation and platform-controlled environments. This requires a fundamentally different mindset.
Rather than viewing agentic commerce as a replacement for traditional marketing, brands should view it as an additional operating layer, one that increases the need for integration across media, commerce, data, and operational functions.
The brands that succeed will be those that accept the complexity rather than resist it. They will invest in the infrastructures required to feed machine systems effectively, while simultaneously doubling down on the human creativity needed to build emotional relevance with people.
Agentic commerce will not simplify the marketing ecosystem; it will create another one alongside it.
Helen Johnson is the managing director of Capture