Yang Han, CTO of programmatic advertising platform StackAdapt, explores how AI agents are solving marketing’s most persistent workflow challenges, from fragmented data systems to real-time experimentation
What are common workflow problems brands have that AI addresses?
A major workflow challenge is the fragmentation of tasks across tools and teams, targeting, creative, optimisation, and reporting, all managed in silos, and often in separate platforms. AI solves this by acting as a conductor across the entire marketing stack, eliminating inefficiencies where campaigns stall because insights sit in disconnected systems.
Another overlooked pain point is the bottleneck between insights and activation; marketers often know what they should do, but can’t scale changes fast enough. Pulling reports, sifting through data, and analysing it for actionable insights can take humans hours or days, but AI can do this instantly, interpreting data and acting on it. This includes auto-adjusting creative, reallocating budget, or orchestrating omnichannel targeting, ultimately bringing agility into the workflow.
What are the implications of AI agents being able to conduct mass experiments at scale for marketers?
When AI agents can run experiments continuously, such as testing creative tones, placements, or contextual alignments, marketers gain more than just efficiency. They gain living models of consumer behavior that update in real time. This shifts experimentation from being episodic (campaign A vs. campaign B) to systemic learning where thousands of micro-tests reveal nuanced preferences across contexts. For marketers, the implication is that strategy becomes less about guessing what will work and more about curating insights from an always-on learning system.
What does it mean for an AI agent to handle ‘complex decision making’, can you give some examples?
Complexity in marketing decisions isn’t about volume of data, it’s about navigating competing goals and ambiguous signals. For example:
Orchestrating first-party data targeting with privacy-safe contextual insights to balance precision with reach.
Pacing budgets dynamically across channels where demand signals fluctuate daily.
Sequencing creative in a way that adapts to the audience journey, showing different ad types or tones based on intent signals.
These scenarios illustrate complexity because there isn’t one “correct” path, only trade-offs. AI’s strength lies in continuously weighing those trade-offs at a scale and speed human teams can’t match.
When AI agents have the capability to interact with other agents – such as customer agents – what impact will this have for brands and marketers?
As consumer-facing AI tools, shopping assistants, recommendation engines, and even personal data agents become more common, the dynamic shifts from brands communicating at scale to brand agents negotiating relevance with customer agents.
The impact for marketers is threefold: they must design offers and messaging that resonate with both humans and their digital proxies, ensure campaigns are interpretable by consumer agents, and leverage the ability to transmit learnings from one AI agent to another, potentially increasing the sophistication and impact of each tool:
- Success depends on designing offers and messaging that resonate with both the human and their digital proxy, which often requires precision in context, tone, and timing.
- Brands will need to optimise not just for visibility, but for machine-readability of value propositions, ensuring that their campaigns and signals are interpretable by consumer agents making filtering decisions on their behalf.
Can you give specific examples of how AI can improve customer lifetime value?
AI can expand customer lifetime value by anticipating customer needs and tailoring engagement across the lifecycle:
Proactive retention: spotting subtle churn signals (e.g., drop in engagement with certain channels) and triggering contextual interventions.
Content cadence optimization: learning the ideal rhythm of brand touchpoints for each customer, whether they respond best to monthly, weekly, or seasonal prompts.
Lifecycle-aware personalisation: dynamically shifting the mix of creative, offers, and even product recommendations as customers move from acquisition to loyalty, ensuring the right product reaches the right user at the right time over the long term to maximise value.
The result is less attrition and more enduring brand affinity.
What are data challenges marketers face that AI can help to alleviate?
The real data challenge today isn’t having too little, but having data that is fragmented, incomplete, or too slow to activate. AI alleviates this by unifying disparate data streams – from first-party identifiers to contextual signals, into orchestrated targeting strategies, interpreting messy or partial data without requiring perfect inputs – reducing reliance on outdated tracking methods, and making insights actionable by feeding intelligence directly into campaign adjustments rather than stopping at dashboards – this transforms raw, often unusable data into decision-ready intelligence that drives performance.
