In the rapidly evolving landscape of digital marketing, few shifts have been as seismic as the rise of agentic AI. To navigate this new terrain, we’re joined by Milena Traikovich, a leading Demand Generation expert renowned for her data-driven approach to nurturing high-quality leads. With extensive experience in analytics and performance optimization, Milena offers a crucial perspective on how autonomous AI systems are fundamentally reshaping the relationship between brands and consumers. Today, our conversation will explore the tangible revenue impact of AI agents as seen in recent holiday sales, the critical pivot from traditional SEO to the new discipline of Generative Engine Optimization, and the delicate balance of managing risk and trust as brand and consumer agents begin to interact on their own.
During the 2025 holiday season, retailers using branded shopping agents reportedly saw 32% faster sales growth. What specific agent capabilities are driving this significant revenue gap, and what is the first practical step for a brand that feels it is falling behind?
That 32% figure is a jarring wake-up call for many, as it shows we’ve crossed a significant threshold. The capabilities driving this gap go far beyond the simple chatbots of a few years ago. We’re talking about agents that are active commercial engines. During that peak Cyber Week period, they weren’t just answering “Where is my order?”; they were autonomously completing complex tasks like initiating returns or updating delivery addresses, with volume for those actions jumping a staggering 70% over the previous year. On Black Friday alone, that number hit 84%. This frees up human teams and backend systems to handle immense traffic without buckling, directly enabling more sales to be processed. For a brand feeling left behind, the first step is a mental one: stop viewing AI as a digital store directory and start treating it as a personal concierge. The immediate practical action is to audit your current AI tools and identify one complex, high-volume task you can delegate to an agent to prove its value and build from there.
Traffic from third-party AI channels is converting at a rate eight times higher than from social media. What characterizes this high-intent discovery process, and how should a marketer’s content and data strategy change to attract and convert these AI-driven referrals?
The character of that discovery process is fundamentally different, and it’s why we’re seeing such incredible conversion rates. A referral from a social platform is often passive—a user is scrolling and gets interrupted by an ad. A referral from an AI channel, however, is the result of a direct, purposeful query. The consumer has already done the work of defining their need and has delegated the research to their trusted agent. They arrive on your site with a pre-qualified, incredibly strong buying intent. To capture this traffic, your strategy must pivot from persuasion to provision. Your content needs to be structured, factual, and easily parsed by a machine. Think less about catchy slogans and more about detailed spec sheets, transparent sourcing information, and clear answers to complex questions. Your data must be impeccable, as an agent’s decision logic will prioritize brands that provide the most reliable and comprehensive information, not just the flashiest marketing.
With predictions that companies may soon spend five times more on Generative Engine Optimization (GEO) than on SEO, what does this shift mean in practice? How does optimizing for an AI agent’s decision logic differ fundamentally from traditional, human-facing SEO tactics?
The predicted five-fold increase in spending on GEO isn’t just an evolution; it’s a revolution in how we approach digital visibility. In practice, this shift is about moving from influencing human perception to influencing machine logic. Traditional SEO is a visual and psychological game—crafting compelling title tags and meta descriptions to catch a human eye scanning a results page. GEO is an entirely different beast. You are optimizing for an AI agent that doesn’t care about beautiful prose but cares deeply about structured data, accuracy, and authority. The fundamental difference is that you’re no longer trying to rank on a public, human-facing page. You’re trying to become the definitive, most trusted answer embedded within the AI’s core reasoning system so that when a consumer asks their agent for a recommendation, your brand is the logical, inevitable choice.
As brand and consumer agents begin interacting autonomously, the risk of missteps that can damage trust increases. What specific “decision checkpoints” or human oversight models are most effective for managing this risk without sacrificing the speed that makes agents so valuable?
This is the delicate dance every brand will have to master. The sheer speed of agentic AI is its greatest asset, but it also amplifies the impact of any error. The most effective model I’ve seen involves creating tiered “decision checkpoints” based on risk. For low-risk, high-volume tasks like a standard shipping inquiry, the agent should be fully autonomous to maintain efficiency. However, for a high-risk action—say, processing a refund over a certain dollar amount or responding to a public complaint laden with negative sentiment—a checkpoint is triggered. This doesn’t mean a human has to approve every step. It could be a time delay that allows for a manual override or a direct flag that requires human sign-off before the agent proceeds. The key is to map out these potential friction points in advance, building guardrails that protect brand values and consumer trust without bogging the entire system down in manual approvals.
The idea of a brand’s marketing agent using deep institutional knowledge to engage a consumer’s shopping agent is compelling. Can you walk me through a specific example of how this works and how that brand agent might learn from repeated interactions over time?
Absolutely. Imagine a consumer has tasked their shopping agent with finding a high-performance laptop for video editing with a focus on sustainable manufacturing. The agent queries several brands. Many brand agents might just return a list of products. But a sophisticated brand agent, using retrieval-augmented generation, taps into its deep institutional knowledge. It accesses not just product specs, but also internal documentation on its supply chain ethics, its carbon-neutral shipping program, and recent positive reviews from professional video editors. It then presents a tailored recommendation to the consumer’s agent, highlighting the specific features that match the user’s core values. When the purchase is made, the brand agent logs which pieces of information were key to the decision. Over time, through repeated interactions with various consumer agents, it learns that mentioning its “closed-loop recycling program” is a powerful conversion driver for environmentally-conscious buyers. It then begins to proactively surface that information earlier in future interactions, constantly refining its approach with no direct human involvement.
What is your forecast for the agentic era of marketing?
My forecast is that the marketing landscape we see today will be almost unrecognizable in 12 to 18 months. The pace is simply breathtaking. The core change will be the inversion of the traditional marketing funnel. Brands will spend far less energy on broad, top-of-funnel awareness campaigns and will instead invest heavily in being the most authoritative and logical solution when a consumer’s agent begins its highly-specific, bottom-of-funnel query. Success will no longer be defined by the creativity of an ad campaign but by the robustness of a brand’s data architecture and its ability to seamlessly and effectively communicate with other AI systems. The marketers who thrive will be those who embrace their new role: not just as storytellers for people, but as architects of trust for machines.
