From generating campaign routes and storyboards to adapting visuals, localising assets and creating multiple formats, AI is reducing the time marketing teams spend getting creative work into production. Recent data suggests the gains are becoming significant, although faster creation is also creating a new challenge: deciding what is good enough, distinctive enough and safe enough to publish.
For years, speed in advertising was largely determined by the production process.
A campaign brief would move from strategy to creative, then through copy, design, storyboarding, production, editing, adaptations and approvals. A single idea could eventually become dozens of assets, but producing those versions required time, people and repeated handovers.
AI is beginning to change that equation.
Marketing teams are increasingly using generative AI not simply to write a caption or create an experimental image, but across several stages of campaign development. It can help teams explore ideas, visualise concepts, build storyboards, generate images and videos, resize assets and create variations before the final work reaches consumers.
The result is a creative cycle that, in some cases, is moving from weeks to days and from days to hours.
HubSpot’s 2026 State of Marketing research, which surveyed more than 1,500 B2B and B2C marketers globally, found that 86.4% of marketing teams were already using AI in at least some areas. Content creation was the most common extensive use case at 42.5%, followed by media creation at 37.2%. Brainstorming, advertising optimisation and administrative automation were also among the areas where marketers were putting the technology to work.
The time being recovered is becoming difficult to ignore. Around one-third of marketers in the study said AI was saving their teams between 10 and 14 hours each week, while another third reported savings of more than 15 hours. More than nine in ten respondents said AI had improved productivity to some degree.
Adobe’s 2026 AI and Digital Trends research offers a similar picture. Conducted with Oxford Economics among 3,000 executives and practitioners, alongside 4,000 consumers, the study found that 76% of organisations had seen moderate or significant improvements in the volume and speed of content ideation and production because of generative AI. Seventy per cent reported improvements in the ability of non-creative teams to generate content, while 69% saw gains in employee productivity.
The numbers suggest that AI’s role in marketing is moving beyond isolated experimentation. But the bigger change is not necessarily that machines can make advertisements. It is that several small, time-consuming steps between a campaign brief and a finished asset can now happen much faster.
Before the shoot, AI is already doing a lot of work
Much of campaign production happens before a camera is switched on or a finished visual is created.
Teams research audiences, explore creative territories, search for references, prepare mood boards, develop copy routes, create mock-ups and turn early ideas into something that clients and internal stakeholders can evaluate.
This is one of the areas where generative AI is proving immediately useful.
A creative team no longer has to fully develop every possible direction simply to see whether it works visually. Image generators can turn rough descriptions into references. Copy tools can develop alternative propositions. Early campaign worlds can be visualised before the team decides whether they deserve a larger production budget.
Canva’s 2026 State of Marketing and AI research, conducted with The Harris Poll, surveyed 1,415 marketing leaders at organisations with more than 500 employees and 3,547 consumers across seven markets, including India. It found that 97% of marketing leaders were using AI in their everyday creative work.
Interestingly, 41% described AI as operating like a “director” within their team, while 39% called it a “collaborator”.
Those descriptions do not mean AI has literally replaced the creative director. Marketers still determine the objective, positioning and final output. They do, however, show how AI is moving deeper into the creative process.
As Emma Robinson, Head of B2B Growth Marketing at Canva, put it, “AI has changed how marketing gets made, but not what makes it effective.”
The difference becomes clearer when looking at how brands are using the technology.
For a recent Lipton campaign in Latin America, Critical Mass used Adobe Firefly during pre-production to explore visual possibilities and help teams align before the physical shoot.
According to Adobe’s case study, more than 100 storyboards and over 2,000 images were developed using around 500 prompts. AI-generated references were also used during casting preparation, helping the agency finalise casting within a week. The campaign itself ultimately featured real models.
The figures are from a vendor-published case study and are not independent measurements of industry performance. But the example illustrates an important part of AI-assisted production: the final advertisement does not have to be generated by AI for the technology to save time.
It can accelerate everything that happens before the advertisement is made.
That changes how teams can approach experimentation. If producing three detailed campaign routes is expensive, a team may commit early to one. If ten rough directions can be visualised quickly, marketers have more room to compare ideas before spending heavily on production.
From weeks of production to shorter creative cycles
AI is now moving beyond pre-production too.
Lysol offers one of the more detailed recent examples.
The brand worked with BCG and Google to develop broadcast-ready 15-second and 30-second spots for Lysol Laundry Sanitizer within an eight-week sprint. Gemini was used during research and ideation, with hundreds of concepts generated and scored before the marketing team manually selected three directions.
The campaign then moved into synthetic production using Google’s Veo and Imagen.
The process was not simply a case of entering prompts and accepting the results. Unsuitable outputs were discarded and regenerated. The brand also established a responsible AI policy before production and worked with consenting and compensated actors whose likenesses were used to create digital twins.
Lysol reported that the process compressed ideation from weeks to hours. The final optimisation stage took two weeks, while the brand also reported an 80% reduction in cost per asset compared with its conventional production process.
In short-term sales-lift testing, the company said its strongest AI-created asset performed almost identically to its strongest traditionally produced asset.
These results were reported by the brand through Google’s marketing platform, rather than independently verified, so they should not be interpreted as a benchmark for what every advertiser can expect.
But Benoit Veryser, VP of Lysol Global and U.S., highlighted a distinction that is increasingly relevant for marketers: “experimenting with generative AI tools isn’t the same as shipping AI-built creative.”
That difference is important.
Generating an attractive image takes seconds. Producing something that can actually run as an advertisement requires far more: accurate products, usable footage, approved brand elements, talent rights, editing, legal review and consistency across every frame.
TSB has approached the problem from another direction.
Its in-house agency Kindred trained an Adobe Firefly Custom Model using hundreds of approved images of Tiny the Elephant, the bank’s brand character.
According to an Adobe case study published in March 2026, creating new imagery featuring Tiny could previously involve a studio process lasting as long as seven weeks. Certain images can now be created in approximately seven minutes. TSB also reported that teams can test dozens of ideas within minutes, revision time has been halved and its broader connected workflow has reduced production time by 60%.
Again, a seven-minute generated image cannot be directly equated with every seven-week studio production. The comparison nevertheless shows where brands may find practical efficiency.
Recurring characters, environments and visual systems do not always have to be rebuilt from the beginning.
Chime has reported similar gains at a broader campaign level.
The financial technology company’s CMO Vineet Mehra said its in-house creative process was using tools including Midjourney, Runway and Veo 3. The company reported reducing creative campaign turnaround by 60%, from roughly ten weeks to four weeks, without increasing headcount, alongside an approximately 30% reduction in content production costs.
Taken together, the examples suggest that AI’s value is not limited to making one task faster. It can reduce the waiting time between multiple tasks.
When making becomes easier, choosing becomes harder
There is another side to this speed.
If producing one asset becomes easier, producing 20 versions becomes easier too.
A campaign can have different headlines for different audiences, several visual treatments, multiple aspect ratios, shorter edits for social platforms, longer videos for digital channels and localised versions for separate markets.
This addresses a real problem for modern marketing teams. Campaigns rarely live in one television commercial or one print advertisement anymore. The same idea may need to appear across social feeds, search, connected television, commerce platforms, websites, email and retailer environments.
AI makes that multiplication easier.
But making more creative does not necessarily mean making better creative.
Canva’s consumer research found that 70% of consumers said they could usually recognise AI-generated advertising because it felt as though something was missing. Another 87% believed the best advertising still required a human touch.
Marketers themselves appear conscious of the issue. Forty-one per cent of marketing leaders surveyed identified poor-quality AI content, often referred to as “AI slop”, as a significant challenge.
HubSpot’s research provides another warning. 62.7% of marketers said more unique, human-centred content would be required to compete with the increasing volume of AI-generated material.
As UserTesting CMO Johann Wrede said in HubSpot’s research, “The human in the loop is the most important part of any kind of workflow, especially involving AI.”
This is where the definition of creative productivity starts to change.
If AI produces 50 versions of the same idea, the marketing team still has to evaluate those versions. Brand managers must check whether the work is recognisable. Creative directors need to judge whether it is distinctive. Legal teams may need to review claims, likenesses and intellectual property. Someone must confirm that the product shown by the model actually exists and looks correct.
The bottleneck can therefore move from creating options to choosing between them.
This is also why generating more advertisements should not automatically be treated as greater creative experimentation.
Fifty headline variations built around the same proposition are still essentially one idea. AI becomes more useful when faster production allows teams to explore genuinely different creative routes before deciding which deserves investment.
The creative may be faster, but the campaign is not automatic
There is one final distinction marketing teams are beginning to encounter.
Producing an asset faster does not necessarily mean a campaign reaches consumers faster.
Adobe’s 2026 research found that despite 76% of organisations reporting faster ideation and content production through generative AI, 53% still described their content supply chains as largely linear and resource-intensive.
That gap matters.
An image might take minutes to generate, but approvals can still take days. A video can be produced quickly, but its claims still need verification. Localised versions may be generated automatically, but cultural context and translation still require review.
The creative process is only one part of the campaign system.
Boston Consulting Group’s 2026 study of 300 global CMOs reflects this divide. While 96% said AI was driving end-to-end marketing transformation, 42% were still primarily using generative AI to assist employees with individual tasks. Only 8% reported campaigns in which multiple AI agents were operating autonomously.
The more advanced organisations were connecting AI across strategy, insights, briefing, content creation, activation and optimisation while retaining human oversight.
That points towards where creative production may be heading next.
Instead of marketers moving manually between a brief, an AI tool, a design platform, an asset library and an approval system, these stages could increasingly become connected.
An approved campaign brief could inform generation. Brand assets could constrain what the model creates. Product databases could provide factual information. Approved master creative could automatically trigger adaptations for different formats and markets. Performance data could then inform the next round of creative testing.
That does not remove the risks.
Generative systems can still produce incorrect product details, inconsistent branding, unexpected visual elements or generic-looking creative. Questions around copyright, training data, talent likenesses and disclosure remain unresolved in several markets. Increased production volume can also overwhelm review teams if governance does not scale alongside generation.
For marketers, the emerging lesson is therefore less dramatic than the idea of AI replacing advertising production altogether.
AI is making specific parts of creative work substantially faster.
It can shorten the distance between an idea and its first visual form. It can create references without a shoot, produce storyboards before production, reuse approved brand worlds, generate multiple formats and make localisation less dependent on rebuilding every asset manually.
But faster generation does not remove the need for judgement.
As creative production becomes cheaper and quicker, marketing teams may spend less time asking, “Can we make this?”
The harder question will increasingly be, “Is this the idea we should make?”
And in an environment where almost every advertiser has access to the same generation tools, that decision may become more important than the speed itself.
Disclaimer: All data points and statistics are attributed to published research studies and verified market research. All quotes are either sourced directly or attributed to public statements.