There's a version of AI marketing that lives entirely on LinkedIn. In that version, every tool is revolutionary, every workflow is "10x," and every person sharing it discovered ChatGPT eighteen months ago and now sells a course about it.
Then there's the version I actually live in: running a digital marketing agency, serving real brands across India and the GCC, trying to figure out which of these tools genuinely earn their place in client work — and which are expensive distractions dressed up in good marketing.
Over the past year at Clazzo, my co-founder Abhijith and I have quietly folded AI into how we work. Not as a headline. As an experiment. Some of it changed how we deliver. Some of it was a complete waste of money. This is the honest breakdown — the one I wish someone had written before I spent the time and money finding out myself.
The hype problem
Here's what bothers me most about the current moment. The loudest voices on "AI marketing" in India are often the ones with the least actual marketing experience. They learned a few prompts, watched a handful of videos, and reframed themselves as experts almost overnight.
That's not a character attack — it's a structural problem. When the people teaching a subject have never had to make it work for a paying client, the advice drifts toward what sounds impressive rather than what survives contact with reality. And reality, in marketing, is unforgiving: a campaign either moves a number or it doesn't.
Domain knowledge compounds. Tool fluency on its own doesn't. Knowing how to prompt an AI video tool is worthless if you don't understand why a particular brand needs a particular kind of video in the first place. The tool is the easy part. The judgment is the hard part — and that's the part no one is selling courses about.
What's actually working
Now the useful part. Here's where, in my experience, AI has genuinely pulled its weight.
Admin and research, not strategy
The single biggest, most reliable win has been compressing the unglamorous work. Competitor research that used to eat hours now takes a fraction of the time. First drafts of content come faster — though they almost always need heavy rewriting to sound like an actual human with a point of view. The pattern is consistent: AI is excellent at removing friction from the preparation, and useless at replacing the thinking.
AI video — impressive, with a catch
The AI spokesperson tools — the HeyGen and Synthesia category — are genuinely impressive when you're working in English. The output quality has reached a point where it's usable for certain kinds of client content.
But the moment you move into Indian languages, the cracks show. When I tested Synthesia with Malayalam, the gap was obvious: pronunciation drifts, the lip-sync wanders, and the result simply isn't production-grade for a brand that cares about how it sounds to a regional audience. For a country with dozens of major languages, that's not a small limitation — it's the whole game. The tools that crack Indian regional languages properly will own a market the current players are largely ignoring.
AI voice — usable, in English
The voice tools have improved dramatically over the last twelve months. A year ago I'd have told you to stay away for anything client-facing. Today, for English production work, they've crossed into genuinely usable territory. The same regional-language caveat applies — but the trajectory is real, and fast.
AI ad creative — depends entirely on the goal
We've run AI-generated ad creative across a number of real client campaigns this past quarter. The honest finding: for awareness-stage campaigns, AI creative has closed much of the gap with human-designed work. For conversion-focused campaigns, the data still tilts toward human-crafted creative. AI is fantastic at generating variety quickly. It's not yet reliable at generating the specific trust signals that actually move someone to buy.
The honest middle
So where does that leave us? Somewhere far less dramatic than either camp wants to admit. AI hasn't replaced marketers, and it hasn't been the empty fad the skeptics enjoy calling it. It's done something quieter and more useful: it's removed a layer of friction from the work, freeing up time to spend on the parts that actually require a human — strategy, taste, judgment, and the relationship with the client.
That's the story I want to tell here, in public, as I keep going. Not "AI will change everything." Not "AI is overrated." Just: here's what I tried, here's what happened, here's what it cost. If it works, you'll see it. If it fails, you'll see that too.
This is the first of what I plan to make a regular habit — honest field notes from inside a working agency, written for founders, agency owners, and marketers who are tired of the hype cycle and just want to know what's real. No course at the end of it. No paywall. Just the work.
