GUEST COLUMN | The hidden AI tax: why chatting with your Facebook ads account might be draining your budget

GURUGRAM: Ankush Vij is an experienced growth marketer who specialises in leveraging technology and data-driven insights to optimise marketing performance and achieve business objectives. He currently serves as co-founder and vice president, media, at Hashtag Orange, a role he has held since January 2025 following four years as vice president, media, and prior experience as director of performance marketing. Over nearly seven years at the company, he has built a strong track record of strategising and managing media campaigns for more than 50 brands across categories including D2C, BFSI, real estate, fintech and gaming.

The rise of AI in marketing has made campaign management feel deceptively simple. Ask a chatbot to “show my top-performing ad sets,” “pause underperformers,” or “launch a conversion campaign,” and it responds in seconds. For busy marketers, that conversational interface feels like the future. But convenience has a cost. And in performance marketing, that cost doesn’t just sit in a software invoice it shows up in wasted time, token usage, fragmented workflows, and sometimes, media inefficiency.

That is why the real debate in Meta Ads management today is not AI versus no AI. It is conversational AI versus command-line automation or, put differently, chat-first convenience versus workflow-first efficiency.

If we look at campaign operations through a business lens, the total cost of management can be broken into four buckets:

Total Cost = Software Subscriptions + Token Consumption + Human Operational Hours + Media Waste

On the surface, conversational tools such as Meta’s MCP connector look highly attractive. They allow marketers to interact with ad accounts in natural language, without code, and can pull insights, create campaigns, and optimize performance directly inside an AI interface. For agencies and marketers who want accessibility, faster onboarding, and low technical barriers, this is a meaningful leap forward.

But the problem begins when a conversational layer becomes the primary operating system for campaign management.

Every question asked in natural language, every follow-up clarification, every repeated prompt, and every back-and-forth action consumes tokens. Multiply that by multiple account managers, multiple clients, and daily campaign checks, and the token meter quietly starts running. Add to that the operational drag of prompting, reviewing, correcting, and re-prompting for structured outputs, and what initially felt “efficient” can become surprisingly expensive. This is the hidden AI tax.

By contrast, a command-line or script-based approach behaves very differently. Instead of asking the system repeatedly what to do, you define the workflow once and let it run in a structured way. Pull weekly reports, flag pacing issues, identify creative fatigue, pause campaigns below threshold, or bulk-edit ad sets across accounts—all without conversational overhead. The output is structured, repeatable, and easier to integrate into internal dashboards or agency systems.

That changes the economics.

First, token consumption drops, because routine tasks are not being re-explained in plain English every day. Second, human operational hours are reduced, because the system runs predefined actions rather than waiting for someone to prompt, inspect, and intervene. Third, and perhaps most importantly, media waste can come down. When alerts, checks, and optimization rules are automated at speed, issues such as overspend, audience fatigue, broken signals, or lagging ad sets are caught faster. In performance marketing, speed is not just an efficiency metric, it is a budget protection mechanism.

This does not mean conversational interfaces have no role. They are excellent for exploration, quick diagnosis, onboarding teams, and making ad platforms more accessible to non-technical marketers. But if the goal is to build a scalable operating model for serious campaign management, conversation should be the front door, not the engine room.

The future of ad operations will likely use both: conversational AI for discovery and decision support, and command-line automation for execution at scale.

Because in advertising, the cheapest click is not the one you buy at a lower CPC. It is the waste you prevent before it is spent.