In the rapidly shifting landscape of 2026, Anastasia Braitsik stands as a definitive voice at the intersection of data-driven results and creative strategy. As a global leader in SEO and content marketing, she has navigated the transition from theoretical AI discussions to a world where automation is a daily operational reality for every major agency and brand. We recently caught up with her to discuss how the industry is moving past the “AI washing” era to focus on tangible business outcomes, such as incremental sales growth and market share expansion.
The following discussion explores the breakdown of human-machine collaboration through the 80/20 rule, the structural changes required as search behavior moves from product-specific to task-oriented, and the rising “agentic” economy. We dive deep into the discrepancies between platform-reported metrics and actual corporate revenue, the risks of autonomous bidding agents, and the future of hyper-personalized, visual consumer interactions.
While AI handles operational workflows and data analysis, human ingenuity remains responsible for defining final outcomes. How do you distinguish between repeatable tasks suitable for automation and the “human spark” needed for strategy, and what specific guardrails prevent AI from optimizing toward gameable, superficial metrics?
In our current environment, we look at the division of labor through a strict 80/20 model where the machine handles the heavy lifting of data manipulation and spreadsheet management that used to consume entire teams. We delegate repeatable tasks—like keyword bidding or real-time budget shifting—to the algorithms because they can reach answers exponentially faster than a human ever could. However, the “human spark” is what we use to determine if those fast answers actually make sense within a brand’s unique historical context. You cannot rely solely on historical data to invent a future that your competitors haven’t already thought of; that requires a creative leap that AI simply isn’t built for yet. To prevent the system from chasing superficial metrics, we have to be incredibly disciplined about what we define as the “outcome.” If you don’t set the guardrails to focus on actual business goals, an AI will become exceptionally proficient at optimizing for a gameable measurement, like a high click-through rate, which looks great in a report but doesn’t actually ring the cash register.
Consumer search behavior is shifting from specific product queries toward task-oriented and problem-solving inquiries. What architectural changes must agencies implement to coordinate intelligence across multiple platforms without wasting budget, and how does this transition disrupt the traditional, linear customer journey that marketers have mapped for decades?
The days of the color-coded, linear customer journey are officially behind us because people are no longer searching for “Product X”; they are searching for a solution to a problem, like “how do I fix a leaky faucet in an old house.” This shift requires a fundamental architectural change in how we manage media, moving toward a system that looks like the center lever on a soft-serve ice cream machine—blending efficiency with bona fide insights. We have to coordinate intelligence across diverse platforms so that we aren’t forcing each individual algorithm to relearn the same campaign every time a user moves from a search engine to a social feed. If we don’t implement this cross-platform coordination, we end up in a cycle of wasted budget where the same consumer is being targeted as if they are at the start of the funnel every single time they switch apps. Agencies now have to build “intelligence layers” that sit above the platforms to ensure that our strategy remains cohesive even as the user’s path becomes increasingly erratic and problem-focused.
Performance marketing often faces a “math doesn’t math” problem where platform-reported growth far exceeds actual corporate revenue. How can brands leverage AI to prioritize incremental sales over platform-specific KPIs, and what steps should a CFO take to reconcile these conflicting data points during earnings calls?
We have all seen the scenario where every platform claims credit for the same conversion, leading to a report where Meta, Google, and Amazon all say they drove 30% more sales year-over-year, yet the CFO sees that actual corporate revenue only grew by 3%. To solve this, we are moving away from the metrics that platforms make available by default and using AI to optimize toward meaningful, incremental business goals. For example, brands like Hershey’s have successfully used these models to prove actual incremental sales during high-stakes periods like Halloween, and Bayer’s Claritin has seen genuine increases in allergy-market share rather than just vanity clicks. A CFO needs to demand a “single source of truth” that uses AI-driven attribution to strip away the double-counting and focus on the bottom line. During earnings calls, the conversation must shift from “cost-per-click” to “incremental revenue generated per dollar spent,” which is the only way to reconcile the inflated platform data with the reality of the balance sheet.
The industry is moving from basic automation toward agentic AI that can learn and self-act. What are the practical risks regarding liability and suitability when deploying autonomous bidding agents, and how can teams “babysit” these systems to ensure they don’t take unintended actions while operating independently?
The move toward agentic AI is like the technology industry’s version of adding truffle oil to french fries—everyone is doing it, but not everyone is doing it well. The real risk with autonomous agents is that they don’t just follow instructions; they start to self-act based on what they learn, which can lead to “suitability” disasters if an agent decides to place a high-value bid on an inappropriate site. We’ve joked about having to “babysit” these systems so they don’t accidentally try to acquire a competitor while we aren’t looking, but the reality of the regulatory filings and liability is quite serious. Teams must implement “kill switches” and strict parameters that define the sandbox the agent is allowed to play in, ensuring it doesn’t wander off-course. While we are moving toward a world of autonomous execution, we are currently in a transition phase where humans must remain the “supervisors” who verify that the machine’s self-determined actions still align with the brand’s ethical and safety standards.
Personalization is evolving toward agent-to-consumer interactions, such as AI chatbots providing specific skincare or cosmetic advice based on user-uploaded photos. How will this level of visual data integration change the way brands communicate, and what must be done to ensure these hyper-personalized recommendations align with a brand’s long-term vision?
This shift toward visual, agent-to-consumer interaction is a massive leap from the text-based chatbots of the past. Imagine a customer uploading a high-resolution photo of their face and receiving a personalized recommendation for microcurrent treatments, blush placement, and a specific serum all in one go. This level of intimacy in data sharing changes the brand communication from a broadcast model to a consultation model, making the brand feel more like a trusted advisor than a faceless seller. However, the danger is that the AI might recommend a product or treatment that is effective but contradicts the brand’s core “look” or long-term vision. To prevent this, marketers must ensure that the AI is trained not just on dermatological or technical data, but on the brand’s specific aesthetic and philosophical guidelines. Every hyper-personalized recommendation must be filtered through a lens that asks, “Does this advice strengthen the brand’s purpose, or is it just a generic solution?”
As digital assistants begin making purchase recommendations on behalf of shoppers, marketers must learn to influence both humans and their agents. What specific tactics are required to maintain discoverability in an agent-to-agent economy, and how do you prevent your brand’s value proposition from being lost in the machine-led translation?
In an agent-to-agent economy, we are no longer just advertising to a person with emotions; we are advertising to a digital assistant that is programmed to find the best value, the fastest shipping, or the most compatible product. To maintain discoverability, our brand data must be structured in a way that is perfectly “digestible” for these machine agents, focusing on technical specifications and verified performance data that an algorithm can rank. However, the risk is that our brand’s unique value proposition—the emotional reason why a human chooses us—gets lost in the machine translation. We have to create a “dual-track” marketing strategy: one track that provides the cold, hard data for the agents to process, and another that maintains the human-centric storytelling that influences the person who ultimately gives the agent its “marching orders.” If we focus only on the machine, we become a commodity; if we focus only on the human, we might never be discovered by the assistant doing the initial research.
What is your forecast for performance marketing?
I forecast that performance marketing will undergo a “great cleansing” where we finally separate genuine, value-driven results from the “AI washing” that has plagued the industry over the past few years. We are going to see a world where the most successful brands are those that have mastered the “agent-to-consumer” relationship, providing such high-utility personalized experiences that the marketing feels like a service rather than an interruption. We will stop talking about “clicks” entirely and instead focus on “agent-influenced market share” and “autonomous incrementality.” The winning agencies will be the ones that have successfully integrated the 80/20 model, allowing their human talent to move away from the spreadsheets and back into the realm of high-level strategy and creative disruption. Ultimately, performance marketing will become less about out-spending the competition and more about out-thinking them through better data orchestration and a deeper understanding of the new, non-linear customer journey.
