Walk into any creative floor today, from a network agency in Bandra Kurla Complex to a boutique shop in Koramangala, and the toolkit looks strikingly similar. The same large language models draft the first ten lines of copy, the same generative platforms spin out storyboards and static variations, and the same AI-assisted edit suites compress weeks of post-production into an afternoon. Access, once the great differentiator between a network giant and an independent, has quietly stopped being one.
Globally, the shift has moved from experimentation to default. Nine in ten marketing agencies now use generative AI in daily workflow, and roughly half have already begun deploying agentic AI to execute tasks rather than merely assist with them, according to Forrester’s State of AI Inside US Marketing Agencies study, released in partnership with the 4As in mid-2026. Tellingly, the heaviest documented use sits in the least glamorous corners of the job, with 74% of agencies using generative AI to summarise documents and 70% applying it to research and competitive intelligence. The creative leap itself is not where the tool does most of its work, a pattern Anthropic’s own Economic Index echoes globally, with creative and media tasks accounting for a little over one in ten everyday AI conversations, and usage leaning more toward augmenting a person’s work than replacing it.
India is following a similar trajectory, even if the domestic numbers are still catching up to a formal count. The Pitch Madison Advertising Report 2026 flagged artificial intelligence and evolving agency structures among the central themes of debate for the year, as the country’s ₹1,55,105 crore ad market grew 12% over the previous year. That figure, and the pace of AI adoption it implies, is worth a desk check before print, since India-specific usage data at the agency level is still thin on the ground.
If the tools are converging, the obvious question follows: when a junior copywriter and a chief creative officer can summon the same hundred headlines in the same ninety seconds, what actually separates one agency’s output from another’s? Three senior voices from the creative and strategy side of the business converge on a surprisingly consistent answer, even as they arrive there by different roads.
The idea still leads
Every technological leap the advertising industry has absorbed, from desktop publishing to digital editing suites, has changed how an idea gets made, not what makes an idea worth making. AI is being read through that same lens by the people running creative departments, less as a rupture and more as the latest, fastest brush in a very old toolbox.
“For agencies, creativity has always been paramount. If everyone has access to the same AI tools, the real differentiation will come from the strength of the idea and how well it connects with people, how naturally it connects with the brand, and whether it can create a meaningful connection or even bring about a change in society,” says Gurbaksh Singh, Chief Creative Officer and Chief Innovation Officer at Dentsu Creative. “The advantage will belong to the agency that can develop the strongest ideas and use technology to make them stand out and come to life most compellingly.”
That framing puts the tool firmly in its place, as an accelerant rather than a source of the idea itself. What has changed is the cost of testing an idea. Where a creative team once had to commit to a single direction before production began, AI now lets them generate and discard dozens of directions before anyone commits a rupee to a shoot. The volume of raw material has exploded, but the volume of genuinely good material has not moved nearly as fast, and that gap is where the argument for human differentiation sharpens.
Judgement as the new scarce resource
A landmark WFA and LIONS study of more than 160 senior marketers and 86 brand owners, representing $141 billion in annual ad spend, found that consistent creative excellence remains out of reach for most global marketers even as AI adoption accelerates, a finding worth verifying against Indian client data before it is treated as a domestic truth. The pattern it describes (more production without proportionately more distinctive work) is exactly what agency leaders are beginning to name as the real fault line of this AI moment.
“As the cost and speed of production continue to fall, the ability to make the right creative choices becomes even more important. AI can generate thousands of combinations and permutations, but it cannot go deep enough to determine which idea is culturally relevant, emotionally resonant in the truest sense,” Singh explains. “When the ability to produce becomes almost limitless, judgement becomes the scarce resource: knowing what to pursue, what to reject, what to challenge, and what to refine.”
For independent agencies, the calculus is the same, only sharper. They now have access to the same generation tools as the big networks, so the old excuse that a smaller team simply cannot produce enough no longer holds. What replaces it is a harder test: can that smaller team consistently tell a merely fluent option from one that is actually good?
“The real source of differentiation, therefore, is the ability to consistently and quickly identify great work from a sea of creative possibilities,” says Luke Dewars, Co-founder and Chief Creative Officer of Goodness of Design and Digital, an independent creative and digital design agency. “Do not spend eight hours writing a single line when AI can help you arrive at options much faster. But deciding whether that line is actually good still requires human judgement. Agencies have been doing that since network agencies came into existence.”
Dewars reaches for an old computer science acronym to describe what happens when that judgement is switched off. Garbage in, garbage out was the warning his school computer teacher gave students learning to code. Fed into an AI model trained on the internet’s existing volume of advertising, the same principle produces a subtler failure mode, one that looks like output rather than garbage at first glance.
“With AI, it can just as easily become ‘clichés in, clichés out’,” Dewars says, before turning to what a model cannot do. “An AI model can scan millions of reviews, identify patterns, and consolidate what people are saying. But it does not experience fear, aspiration, humour, or vulnerability, and it carries no accountability for the creative decision. A human understood fear and framed, ‘Darr ke aage jeet hai.’ Other humans then assessed whether that thought could connect with an audience and speak to the masses.”
Where AI actually earns its keep
That Forrester split is worth pausing on, because it complicates a lazy version of the AI disruption story doing the rounds at industry panels. The heaviest, most confident use cases (summarising brand guidelines, pulling together a competitive scan before a pitch) sit squarely in the pre-creative and post-creative layers of the job. Freeing that time is not a small thing, but it is not the same as freeing an agency from the need for a distinctive idea, and conflating the two is the mistake several leaders interviewed for this piece are implicitly warning against.
Taste, strategy and the human account
If Singh and Dewars locate the differentiator inside the creative department, the strategy side of the business is making a parallel argument about what happens upstream of the brief. A model can answer a question with impressive speed. It still needs a human to decide which question is worth asking in the first place.
“AI will increasingly level the production playing field, but access to the same tools does not create the same thinking. Differentiation will come from the quality of the question, the strategic insight behind the idea, cultural understanding, and the judgement to know what is worth making,” says Ambika Sharma, Founder and Chief Strategist at Pulp Strategy, a Delhi-headquartered full-service digital, experiential and technology-led marketing agency. “AI can generate hundreds of options. The advantage lies in knowing which one matters and why.”
Sharma’s second point pushes the argument further, separating competence from distinctiveness in a way that maps onto Comcast Advertising’s 2026 State of AI in Advertising survey, where a majority of senior agency and brand buyers globally said AI had not yet produced meaningful business results, despite near-universal belief that it was transforming the category. Faster output, in other words, has not automatically meant better output, and that gap is precisely where taste is meant to do its work.
“As production becomes faster and cheaper, judgement becomes more valuable, not less. Taste is the ability to recognize what is distinctive rather than merely competent,” Sharma says. “In fact, as AI increases the volume of competent work, originality and judgement become scarcer, and therefore more valuable.”
The word competent is doing a great deal of work in that sentence. It describes exactly what a well-prompted model now produces on demand: correct, on brief, unobjectionable, and largely indistinguishable from what every competing agency could produce with the same prompt. Advertising has rarely rewarded competence. It has rewarded the one idea that made a boardroom lean forward, and that idea has never come from the model alone.
If judgement is genuinely the scarce resource everyone agrees it is becoming, the more interesting question for agency leadership is whether it can be built deliberately rather than left to chance. The WFA and LIONS Clients and Creativity 2026 study argues that brands seeing the most consistent creative output are the ones that have codified creativity into a system, with clear review gates and ongoing investment in talent and skills, and AI assigned a clearly bounded supporting role rather than free rein. Translated to an Indian agency floor, that could mean a senior creative sign-off built in at the exact point where AI-generated options multiply fastest, and treating the ability to spot one original line among fifty competent ones as a skill worth teaching junior teams deliberately, rather than assumed to arrive with experience alone.
Even Cannes Lions, the industry’s own arbiter of great work, spent its 2026 edition wrestling with this exact tension. The introduction of a dedicated AI craft category reopened an old argument about who, or what, gets to decide that AI-assisted work is genuinely excellent, with several jurors publicly questioning whether machine-made craft should be judged by human instinct at all. That the industry cannot yet agree on the answer is telling. It suggests the debate over where AI’s contribution ends, and human judgement begins runs through the institutions that set the benchmark for the entire business, not just individual agencies.
What emerges from these conversations is not a rejection of the technology, but a recalibration of where value sits inside the agency model. Production, once the scarce and expensive part of the process, is becoming cheap and abundant. Judgement, the ability to look at a hundred AI-generated options and know within seconds which one is worth a client’s trust, is becoming the part that cannot be commoditised. Agencies that treat that judgement as a skill to be deliberately built, hired for, and protected, rather than an accident of having good people on staff, are the ones likely to hold their ground as the tools keep getting easier for everyone else to use.
Leave a Reply