For years, AI in adtech meant optimization: smarter bids, tighter targeting, better performance. But that’s changing. We’re entering AI’s second wave, where systems don’t just optimize execution, they make decisions in real time, across channels, platforms, and customer journeys. This shift isn’t about speed alone. It’s about intelligence, autonomy, and orchestrating the entire funnel with purpose.
For the past years, AI in adtech has been largely associated with one thing: performance optimization: think better bidding, tighter targeting, better performance. But a new chapter is unfolding.
Welcome to the second wave of AI in advertising: where machines aren’t just tweaking dials, but making decisions in real time and across platforms, channels, and customer contexts.
And this time, it’s not about speed. It’s about intelligence, autonomy, and full-funnel orchestration.
The first wave of AI delivered undeniable value. Machine learning models helped brands bid more efficiently, segment audiences faster, and run A/B tests at scale. Yet, in traditional adtech stacks – even though powered by AI – intelligence was fragmented. Demand-side platforms (DSPs) optimized bids. Supply-side platforms (SSPs) optimized yield. Customer data platforms segmented audiences. Creative tools tested variations. The list goes on.
But these tools largely lived inside silos – each system had its own logic, its own data, and its own KPIs.
The second wave of AI aims to change this model. Instead of optimizing individual components, AI will increasingly act as a decisioning layer that connects them all.
In contrast, the second wave of AI is breaking those silos. Today, AI systems can decide: Which audience to target? Which channel should be used for it? Which creative would work best for that specific audience? And, what’s the best ‘value for money’ option in that specific ad setting?
How? The new AI systems can track and instantly process real-time signals flowing across DSPs, SSPs, CDPs, ad servers, and analytics platforms.
This move toward platform-level intelligence is no longer theoretical. According to a 2025 Gartner survey, 65% of CMOs believe AI will dramatically change their role in the next two years, as decision-making increasingly shifts from isolated execution to AI-orchestrated systems spanning the full marketing stack.
Several factors are converging now to make this second wave inevitable.
1. The explosion of first-party data
With third-party cookies on the verge of being phased out, not only from a technical perspective but also from a belief standpoint, brands, retailers, publishers, telcos, and banks are sitting on increasingly valuable proprietary data. AI is needed to interpret and activate this data at speed and scale.
2. The rise of real-time channels
CTV, retail media, and in-app advertising demand decisions in milliseconds. Static rules and manual optimization simply cannot keep up.
3. Advances in AI infrastructure
Cloud-native data platforms, streaming pipelines, and modern machine learning (ML) frameworks now allow models to learn and act continuously, not just in batch cycles.
McKinsey recently explained this shift in how organizations are moving beyond basic AI use toward agentic AI systems capable of acting autonomously across business functions – showing the trend from simple analytics to AI making decisions across entire workflows.
In this second wave, AI does more than optimize bids.
It dynamically reallocates budgets across channels. It predicts which creative will perform best for a specific user context. It adapts frequency and sequencing across touchpoints. It balances short-term performance with long-term customer value.
And all these decisions are not made in isolation. They are informed by:
• CDP-level customer profiles.
• Privacy and consent signals.
• Inventory and supply-side constraints.
• Business rules and brand controls.
The result is not full automation without oversight, but AI-assisted autonomy, where humans (still) define objectives and guardrails, and our beloved machines take care of execution at scale and speed.
This second wave requires a different kind of adtech architecture.
AI can no longer live only inside a DSP or recommendation engine. It must sit above and across the stack, acting as a shared decisioning layer.
That typically means, among other things, these issues are still being figured out:
• Event-driven, real-time data pipelines.
• Unified identity and consent frameworks.
• Interoperable DSP, SSP, and ad server components.
• ML models embedded into core workflows, not bolted on.
• Explainability and governance built into decision logic.
In other words, AI becomes part of the platform’s backbone, not just a feature.
Many off-the-shelf adtech platforms were designed for the first wave of AI. They optimize within their own silos, using limited data and fixed logic, and this can lead to truly lamentable consequences. One very popular brand neglected the ‘old stack limitations,’ and its ads ended up being shown next to harmful content.
Hence, as companies push toward real-time, cross-channel decisioning, these limitations become clear:
• Data cannot move freely across systems.
• Models cannot be customized to specific business goals.
• Decision logic is opaque and hard to govern.
• Innovation cycles are constrained by vendor roadmaps.
This is why we increasingly see enterprises exploring custom or hybrid adrech stacks, where AI decisioning can be tailored to their data, channels, and regulatory context.
The second wave is about control, not just performance. The real promise of AI’s second wave in adtech is not just better performance metrics. It’s control.
Control over data. Control over decision logic. Control over how AI operates within regulatory, brand, and ethical boundaries.
By 2026, AI-driven decisioning will be a baseline expectation in advanced adtech environments, especially in retail media, CTV, and data-rich enterprises.
The winners will not be those who simply adopt more AI tools, but those who design coherent systems where AI can learn, decide, and act across the entire advertising lifecycle.
The second wave has already started. The real question is whether your adtech stack and roadmap are ready to support it.
