In a world where AI and data are reshaping marketing, Varun Seth is helping enterprises turn complexity into intelligent, automated decision-making. As Managing Partner at AGL – Hakuhodo and head of the agency’s AI-first consulting practice, dXfactor, he drives Martech transformations that blend strategy, technology, and execution at scale. With over 15 years of experience, Varun has led large-scale digital transformations, proprietary product innovations, and full-funnel marketing strategies across industries including automotive, consumer electronics, real estate, and telecom.
A thought leader in AI-driven marketing, he specializes in creating connected, intelligent ecosystems that enable real-time decisioning, optimize ROI, and scale operations efficiently. Varun combines technical expertise in computer science with business acumen, ensuring that complex Martech architectures deliver practical outcomes. In this conversation, he shares insights on AI-led marketing acceleration, building resilient operating models, and how enterprises can harness data and automation to stay ahead in a rapidly evolving digital landscape.
Agentic AI is often talked about as automation at scale. In your view, where does it start influencing judgment rather than just execution?
The shift from execution to judgment begins the moment we move from rules to goals. Traditional automation is deterministic, as it follows linear if-this-then-that logic. Agentic AI, by contrast, evaluates competing priorities and makes trade-offs rather than simply executing predefined instructions.
It weighs growth versus efficiency, short-term performance versus long-term value, and signal strength versus risk, and it recommends actions based on probabilistic outcomes rather than certainty. At that point, AI is no longer just accelerating work; it is shaping what gets done and why. The shift is from doing things right (efficiency) to doing the right things (effectiveness).
For years, MarTech has been built around execution: campaigns, channels, dashboards. What’s fundamentally changing now that makes decision-making the new focus?
We’ve reached a point of execution saturation. Teams have more data, channels, and tools than ever, but also far greater complexity. What’s fundamentally changed is that latency is now the enemy of ROI. By the time insights are reviewed and decisions approved, opportunities have often passed.
As a result, marketing is shifting from retrospective reporting to predictive decisioning. The modern MarTech stack is no longer just about how you reach customers, but how quickly and
intelligently you decide why to reach them, evolving from an execution toolkit into a decisioning layer focused on business outcomes, not just activity.
Most enterprises already struggle with fragmented MarTech stacks. Does agentic AI simplify this problem, or does it raise the bar for how MarTech architecture needs to be designed and governed?
Agentic AI simplifies the user experience but raises the bar for architecture. It does not automatically fix fragmented MarTech stacks, it exposes their weaknesses. AI systems depend on clean data flows, clear ownership, and strong governance. Without these foundations, intelligence becomes inconsistent or misleading. You cannot layer intelligence over chaos.
Enterprises therefore need to design MarTech architectures with intent: clarity on decision ownership, interoperability across systems, and governance models that define how AI recommendations are validated, overridden, and acted upon. When done well, agentic AI can dramatically reduce surface complexity for teams, but only if the underlying architecture is built for intelligence, not accumulation.
This is precisely the gap many enterprises are now closing through integrated operating models—such as the one we drive at Dxfactor. We approach MarTech as an end-to-end ecosystem rather than a collection of disconnected tools. By aligning go-to-market strategy, architecture, and measurement within a single ‘decision framework,’ organisations can finally turn AI recommendations into accountable actions.
“The shift underway is not about adding more intelligence; it is about designing marketing systems that are mature enough to absorb, govern, and act on that intelligence.”
How does the rise of agentic AI change the relationship between strategy, media execution, and measurement, areas that have traditionally operated in silos?
Historically, strategy, media execution, and measurement operated in a linear relay: slow, sequential, and siloed. Agentic AI collapses this into a continuous intelligence loop. Strategy becomes dynamic guardrails rather than a static plan, measurement happens in real time at the point of execution, and media adjusts instantly based on live signals.
When the same system executes and measures outcomes, silos disappear. Marketing moves toward autonomous operations, where decisions, actions, and learning happen simultaneously, and strategy is continuously refined by real-world performance.
Are CMOs today equipped to trust AI-led decisions, or is the bigger gap cultural and organizational rather than technological?
Technology has outpaced organisational culture. The bigger gap today is not technological, but organisational. Most CMOs already rely on AI-driven systems, often without fully realising it, yet lack a clear operating model for trust and accountability.
Trust in AI-led decisions depends on clarity: who owns the decision, how recommendations are challenged, and where human oversight applies. Without this, AI remains a black box. Organisations that succeed treat AI not as a replacement for leadership judgment, but as a decision partner governed by clear rules and accountability.
By 2026, what decisions do you believe marketing leaders will no longer make manually, and why?
By 2026, marketing leaders will step away from high-frequency, data-intensive, and reversible decisions. Humans are not built to make hundreds of micro-decisions consistently every day.
Instead, leaders will focus on defining objectives, constraints, and success metrics, while agentic systems handle continuous optimisation within those guardrails.
“Leadership shifts from making every decision to designing the system that makes decisions well.”
