The AI train has left the station. Are Indian agencies on it?

AI and advertising have taken up a fair bit of space in recent times. Look at this in the context of brand-name agencies in a hurry to delete their legacy, the emergence of new alphabet-based shops, and staff cuts across the world. Enter the AI express: did agencies get the window seat to admire tomorrow’s view, or are they hanging on like rush hour commuters in a Mumbai local?

The hustle is real. The big global networks have pledged major investments. The demo and results of platforms like WPP Open are fast and impressive. But when a system can churn out hundreds of options in minutes, the challenge is no longer creation. It’s judgment. It takes a creative director’s instinct to say, ‘Yes, it can be done, but should it?’

In the indie agency world, a few ChatGPT prompts, a Midjourney moodboard there, a track from Suno and ten PowerPoint slides later, the deck proudly declares, ‘AI-enabled agency.’

The reality: it’s icing on a pre-2023 cake. Perhaps it’s time to rebuild the oven.

Are machines getting the message?

Good briefs still struggle to find that elusive insight. Yet somewhere along the journey, the agency is caught between emotion and render. Usually, this means coaxing polite prompts (yes, AI has feelings too) to create last-minute visuals. AI, like any good creative, has good days and otherwise. When it delivers, it is on fire. When it isn’t in a mood, you can hear the sulk.

One of our clients pointed us to Coca-Cola’s ‘New Guy’ campaign film during a briefing. We loved the reference and went down to the frame break-up.  At the core of the film, there was chaos, confusion, and connection. Capturing that in a prompt wasn’t easy. Yet it was necessary. AI filmmaking gets richer as we delve into the director’s mind.

Film director Mani Ratnam, for instance, often uses bus journeys when there’s more telling than showing -- a moving backdrop keeps emotion visible. We used this thinking while creating an AI-generated Valentine’s Day film for a jewellery brand. These are lessons from legends who live craft, not from prompts. AI can copy, but it still needs a human prompt to lift it

Real progress would mean that every brief starts with ‘How can AI spark human insight?’ Without that intention, the process turns repetitive fast.

In one project for a public–private infrastructure client, the model even surprised us by suggesting Kannada shop signage in the background (we’re Bangalore-based). A small, local detail that instantly grounded the story. Sometimes, even the machine gets the message.

Creativity in a snap, but at what cost?

On the creative side, AI has moved from novelty to normal. Most agencies now have someone fluent in Midjourney, Gemini, Veo3 or ChatGPT. In a world of tool parity, the differentiator goes back to advertising basics--the idea, the script, the eye.

Take Tilly Norwood, a virtual actress created by AI talent studio Xicoia. When Norwood’s first acting part--a single AI-generated comedy sketch called AI Commissioner—was released, she wrote on Facebook, “I may be AI-generated, but I’m feeling very real emotions right now.”

The reaction from Hollywood was quick: “That’s AI? Good Lord, we’re screwed.” Soon after, Vogue’s August issue featured AI-generated models, reopening the debate over what’s real, what’s not, and what’s AI slop.

 

 

 

 

View this post on Instagram

 

 

 

 

 

 

 

 

 

 

 

A post shared by Tilly Norwood (@tillynorwood)

Big networks build proprietary platforms. Independent agencies move faster, burn through tool credits and thrive on curiosity. It’s not capital that sets the pace; it’s hunger. The process is now the real difference. One side builds systems, but the other hits render faster. Both are ‘AI-enabled,’ just at different speeds. Who has better continuity in AI creation? Scalability? Chances of avoiding burnout?

Agencies are learning on the go, discovering new tools almost every week. The change is so quick that most are improvising. There are moments of ‘Klueless in Koramangala’ panic. And this year’s Cannes entries were a stark reminder of the cost of speed. Award-winning entries were pulled, rules tightened, and regulation quietly became the headline.

Tech vs creative comes down to pushing past friction

Today, agencies are part creative studio, part production house, and part tech lab. Dedicated AI festivals now list concept, pre-production and post-production as entry parameters, asking that at least one AI tool be used at each stage. On some jury panels, CCOs are joined by tech leaders who can see both the art and the algorithm.

ByteDance and Alibaba may not come from agency land, but their models understand human behaviour better than most of us care to admit. The Diwali flood (slop?) of AI-generated videos proved one thing: volume is easy, recall is hard. A Chinese algorithm can pick a colour palette that “boosts engagement,” or an American model can spin a catchy hook, but when ten brands post the same festive-looking reel, “made by AI” isn’t much of a story.

So the real question is: do we drive change, or do we just consume the tech?

Agencies should nudge toolmakers to build for storytellers, not the other way around. Platforms like Adobe Firefly, GenStudio and Canva have already given clients the same creative power. Access is not the advantage anymore.

The real friction isn’t in using the tool, but in how we push it through culture and behaviour. That moment, when human instinct meets machine output, is an agency’s real DNA. The creative director’s job is knowing when to add a little friction, to push AI past its comfort zone. You could go with the rave reviews that AI gives you, or you could tend to that spark until it produces something original.

When Channapatna toys wanted to tell a Bengaluru luvvu (love) story, it took several prompts to move from LLM to ‘Local Language Model.’ Similarly, when working on AI videos for overseas markets, the tiniest details matter. In different cultures, who sits where at a family dining table? Details like these warrant a human touch.

The speeding AI train

Understandably, clients often have their calculators out. Most compare AI work to traditional production costs. “How much cheaper is it?” is often the first question. Experimentation budgets are small, and quality expectations are flexible because, well, “It’s AI, it’s okay.” Larger agencies often outsource AI projects to smaller, faster shops or independent creators. Everyone’s learning to balance speed, quality and cost.

Today’s creative is expected to think like a planner, code like a techie, frame like a DOP and juggle timelines like Christopher Nolan. The real skill now is adaptability, being ready to experiment, fail and adjust. Hiring can’t depend on ‘five years of Photoshop’. Until new roles take shape, curiosity and openness to change are still the best qualifications.

But the fact that the AI train doesn’t wait for perfectionists should suit us agency folk just fine. We’re always a work in progress, fixing slides in the client’s elevator and rewriting briefs over lunch. If we hesitate to experiment with AI now and learn by “failing fast”, we’ll end up on the platform, waving as the future speeds away.

Instead, climb aboard the AI train and hang on tight -- Fevicol style.

-Niranjan Natarajan, founder and creative director, Why Axis Advertising