There is an enormous amount of hype surrounding AI in adtech right now, but practical, real-world examples of successful implementation can be difficult to pin down. On this episode of The Yield Doctor podcast, host James Deaker sits down with Nagarajan (Naga) Chakravarthy, Chief Digital Officer and founding management member at IOPEX, to cut through the noise.
IOPEX has spent the last 15 years helping global retail media brands and agencies optimize operations. Today, they are at the forefront of AI transformation for digital advertising revenue operations.
Here is a breakdown of how AI is shifting from theoretical hype to everyday operational reality in the advertising landscape [00:00].
The Evolution: From SaaS Workflows to “Command Agents”
For the past decade, software-as-a-service (SaaS) companies focused on simplifying workflows for operational workers. AI is changing the paradigm entirely by creating software that acts like a worker itself. IOPEX terms these “Command Agents”—also known as agentic AI [03:04].
Instead of just organizing tasks, these agents perform complex actions directly to streamline processes for human operators.
3 Core Pillars of AI Process Automation
Naga abstracts the current practical implementations of AI into three core functionalities [04:05]:
- Validation & Prediction: Using past data and established guidelines to validate the current state, making future forward-modeling more predictable.
- Next-Best Action: Analyzing historical situations to determine the most successful next action for a current scenario.
- Execution (Agentic AI): Converting these data-backed decision points into immediate, automated actions.
Real-World Use Cases in Ad Operations
Rather than keeping things theoretical, Naga points to specific live implementations and proof-of-concepts making waves today:
- Media Planning Optimization: Ingesting two years of historical campaign performance data, media plans, and insertion orders (IOs) into a vectorized database [13:54]. Users can prompt the AI with a specific budget and target demographic to receive multiple plan variations instantly. The AI handles the cognitive memory load, allowing human users to interact, combine plans, and quickly standardize the output into formal IO templates [07:05].
- Creative and Content Testing: Using AI to generate initial concepts and format resizing [06:07]. While AI can manage the “IQ” part of structural resizing, human judgment and emotional intelligence (EQ) remain necessary to polish the content and make it resonate creatively [16:35].
Key Operational Mindset Shifts
Bringing AI into ad operations introduces critical trade-offs and structural changes that organizations must prepare for:
- Speed vs. Oversight: Human teams are shifting away from executing the literal mechanics of a task to instead managing and overseeing the task executed by the AI [09:06]. A planner who once built 5 complex media plans a week might now oversee the generation of 15 to 25.
- The Skill Shift to Prompting: Value is moving away from knowing the raw nuances of an execution tool toward knowing how to communicate with an AI via prompting to get the best inference [10:14].
- New Error Patterns: Human “silly mistakes” are replaced by system anomalies like AI hallucinations [11:08]. The quality control function must adapt to catch these unique, non-human edge cases.
Misconceptions and Obstacles
Naga emphasizes that two primary misunderstandings slow down AI adoption [19:10]:
- The Expectation of Immediate ROI: AI is not a magic wand [20:39]. Monetary gains arrive over time through modular implementation and parallel runs that let the system adapt safely.
- The Fear of Job Loss: The obstacle isn’t human replacement, but rather a temporary skill gap [25:38]. Workers will not lose jobs if they adapt to using the technology to redefine their roles into higher-level, strategic management positions.
The Impact on Publishers and the Playing Field
Ultimately, Naga views AI as a net positive for digital publishers [29:57]. By leveraging AI to optimize ad placement and bidding in real time, publishers can increase their ad real estate value, boost fill rates, and drive higher bid prices [30:38].
Furthermore, AI serves to level the playing field [28:03]. While it makes multi-million dollar budgets more predictable for enterprise spenders, it simultaneously democratizes insights for small advertisers, allowing them to optimize tight budgets and scale much faster than previously possible.