Key Takeaways
- Fresh Step slashed its ad creation costs by a factor of 10 with AI, going from weeks to hours by using new tools for content and audience targeting.
- Instead of hiring agencies for everything, AI platforms can now generate visuals, write scripts, and even produce voiceovers, cutting out huge chunks of traditional production costs.
- To make AI work, you need clean first-party data. The models need good information to train on for targeting and personalization, which is what actually drives your ROI.
- Marketing teams have to change how they work. The job is no longer about manual creative work but about strategic direction, writing good prompts, and analyzing performance data.
- Don’t try to switch everything at once. Start with a pilot program for a specific campaign to test the AI tools and figure out a workflow before you roll it out everywhere.
You hear a lot about efficiency in marketing, but a real 10x jump in a core process is almost unheard of. Fresh Step just did it, completely overhauling how they make ads using artificial intelligence. Their journey from a brief to a live ad now happens in a tiny fraction of the time and budget, showing everyone what’s actually possible for cost-effective campaigns. So how did they pull off such a massive cost reduction, and what can the rest of us learn from their experience with AI ad creation?
The Creative Bottleneck: Before AI
For as long as I can remember, the old advertising workflow was stuck in a rut. A brand writes a brief, sends it to an agency, and then you wait. The agency goes through concepts, scripts, storyboards, casting, shooting, editing, and final delivery. That whole process was a slow, expensive, linear slog. Even a simple digital ad could take weeks or months, looping through a dozen people for feedback. Every single change request adds to the bill and, more importantly, delays getting your ad in front of customers, meaning you miss opportunities.
Think about the standard drill: Fresh Step wants to push a new product or run a seasonal sale. They brief their agency on who they’re trying to reach and what they want to say. The agency then assigns a team of writers, art directors, and producers who spend a week just coming up with initial ideas. Then you have rounds of internal brand reviews and legal checks that can burn another two or three weeks. After that, you finally get to production, which means booking studios, hiring talent, and all the post-production work. Any of these steps can create delays and blow up the budget. And the problem gets worse when you need a ton of different assets for a campaign (think different sizes for social, different lengths for video, statics for display). The old model produced good work, sure, but it was never built for the speed 2026’s digital market demands.
Fresh Step’s AI Transformation: From Weeks to Hours
Fresh Step tackled this problem head-on by weaving AI into each step of its ad creation workflow. They had one clear goal: slash the time and manual work involved while still producing good, on-brand creative. The results were big. By using generative AI, they took the time from brief to a finished ad from around three weeks to just a few hours for a lot of their campaigns. This involved rethinking their entire creative process from the ground up.
A huge part of the shift was in concept generation and scriptwriting. Instead of waiting for copywriters to stare at a blank page, Fresh Step’s marketing team now plugs their campaign brief, audience data, and past performance metrics into a custom AI model. Because this model was trained on years of their own successful ads and brand guides, it spits out multiple script variations, taglines, and message angles that actually sound like them. The marketer’s job changes from pure creation to smart curation and editing. For example, the AI can generate ten different 15-second video scripts for a new cat litter scent in less than an hour, each one already formatted for platforms like TikTok for Business or Meta’s Ads Manager, and it even suggests visuals. This alone collapses the ideation phase.
Creating visual assets also got a complete makeover. Fresh Step started working with an AI platform that generates photorealistic images and short videos from text prompts. For their product ads, they can upload a 3D model of their packaging and just tell the AI to place it in different settings: “a modern living room with a tabby cat,” or “a minimalist apartment with natural light.” The AI renders these scenes quickly and with surprising quality, giving them dozens of unique images in minutes. This completely gets rid of the need for expensive photoshoots for many standard ad types, which was a huge part of their cost-effective campaigns. Voiceovers, which used to be another time-consuming and expensive step, are now also made with AI. Using synthetic voices that they can tweak for tone and emotion, they can produce multiple voiceover tracks for A/B testing almost instantly.
The Technology Behind the Savings
Fresh Step’s success comes from its smart adoption of a few specific AI technologies. First, they put money into large language models (LLMs) and fine-tuned them on their own brand voice, product details, and historical marketing data. This custom training is what enables the LLMs to write copy that sounds genuinely like “Fresh Step” and keeps the brand consistent. If they hadn’t done this, the generic AI outputs would need so much human editing that it would wipe out most of the efficiency gains.
Second, the visuals are created with advanced diffusion models. These models are good enough now to create very detailed and context-aware images from a simple text description. When the prompt is “a happy cat playing near a Fresh Step litter box in a sunlit kitchen,” the AI generates a unique, high-res image that fits the bill, down to the lighting and textures. This tech has gotten so much better over the last two years that it’s now a serious option for commercial work, not just for abstract art.
Finally, Fresh Step connected these generative tools to their existing martech stack. That includes their digital asset management (DAM), customer relationship management (CRM), and ad serving platforms. This means the AI-generated images and videos can be tagged and pushed directly into platforms like Google Ads or the Meta Business Suite for deployment. This tight integration, where the whole workflow from idea to launch is managed from one dashboard, is what really delivers the 10x cost reduction. It cuts down not only on production time but also on all the administrative busywork of managing a campaign.
Data, Ethics, and Human Oversight
AI can make things efficient, but its actual effectiveness comes down to the quality of the data you feed it. Fresh Step knew their whole project depended on giving their AI models clean, complete first-party data. This meant feeding it detailed customer segments, historical campaign performance (like CTR, conversions, and CPA), and all their brand guideline documents. The more context the AI has, the better it gets at guessing which creative will work for which audience. For example, if the data shows that city dwellers respond better to ads with a minimalist look, the AI can automatically generate visuals and copy that fit that preference.
Of course, the ethical side of using AI in ads is a huge deal. Fresh Step put strict guidelines in place and made sure a human marketer reviews every single AI-generated ad before it goes out. This is for more than just quality control. It’s to check for compliance, catch any unintentional bias, and protect the brand’s voice. AI can sometimes generate weird, inappropriate, or just off-brand content, and you absolutely need a person to catch that stuff. In my experience, brands often want to skip this final human check to save time, but it’s a terrible idea. It always costs you more in the long run. Fresh Step is also very deliberate about its policies on synthetic media, especially when it involves people or pets, and they’re aware that consumer feelings about AI-generated content are still evolving.
Rethinking the Marketing Team’s Role
This shift to AI for ad creation redefines the marketing team’s job instead of eliminating it. The team at Fresh Step has moved from being producers of content to being strategic architects and performance analysts. They now spend their time writing precise prompts for the AI, A/B testing the AI’s output, and digging into performance data to figure out how to make the next set of prompts even better. This requires a different set of skills: you need to understand how these models think, be able to act fast on new data, and keep a firm grip on the overall strategy.
For instance, a campaign manager might spend their morning tweaking prompts for video ads aimed at new cat owners, testing emotional angles like “comfort” against “playfulness.” In the afternoon, they’re looking at the results from those ads to see which visuals or voiceovers got the most engagement. This constant feedback loop is what makes the system so powerful. The AI learns from what works and gets better over time. A 2024 IAB Outlook Report found that almost 70% of advertisers are planning to spend more on generative AI tools, which shows this change in roles is happening everywhere. The future of marketing is about intelligent orchestration, not manual labor.
The Future of AI in Advertising
Fresh Step’s story is a wake-up call for the whole industry. Their 10x cost savings isn’t a fluke. It’s a glimpse of what’s now possible for any brand that’s willing to adapt to these AI tools. Looking at how fast AI is developing, we can expect even more powerful capabilities soon. AI will get better at understanding brand nuances, creating hyper-personalized ads for individuals at scale, and even predicting how well a campaign will do before you spend a dime on media. This is especially exciting for smaller businesses and startups, because it gives them access to high-quality creative production that used to be reserved for huge companies with massive agency retainers. The question for marketers isn’t *if* they should adopt AI anymore, but how fast they can get it integrated into their daily work to keep up.
AI ad creation is quickly becoming a standard part of a modern marketing operation. The brands that get ahead will be the ones that focus on feeding the AI good data, using it ethically, and training their teams for a new way of working. Those are the companies that will define what efficiency and effectiveness look like in the years ahead.
How does AI specifically reduce the cost of ad creation?
AI cuts costs by automating tasks that used to require expensive people and resources. Things like writing scripts, creating images, and recording voiceovers can now be done by AI, which means you spend a lot less on agency fees, studio time, and production crews. It also speeds up the whole process, so time isn’t wasted on endless review cycles.
What types of AI are most commonly used for ad creation?
The main tools are Large Language Models (LLMs) for writing text like scripts and headlines, diffusion models for making images and videos, and speech synthesis AI for voiceovers. Many also use predictive AI to help with audience targeting and to forecast campaign results.
Is human oversight still necessary when using AI for ad creation?
Yes, absolutely. A person still needs to review everything the AI produces. You need to check for brand consistency, make sure it follows advertising rules, and catch any weird or inappropriate outputs. The human provides the strategy and the final stamp of approval.
How can a brand ensure AI-generated ads maintain brand voice and quality?
To keep the brand voice consistent, you have to train your AI models on your own stuff: your brand guidelines, your past successful ads, and performance data. It also requires constant human feedback and tweaking the AI prompts to make sure the output stays on track.
What skills do marketing teams need to develop for AI-driven ad creation?
Teams need to get good at prompt engineering (telling the AI what to do), data analysis (understanding the results), and strategy (guiding the AI’s direction). The job is less about doing the manual work and more about being the conductor of an AI orchestra.