Each AI advertising and marketing catastrophe you’ve seen—the cringeworthy pretend influencers, the clearly robotic copy, the manufacturers caught mendacity about AI use—occurred as a result of somebody thought AI may substitute technique, creativity, or authenticity.
The examples on this article show the alternative: AI’s worth is eradicating grunt work and constraints, not changing the people who perceive your market, your model, and your clients.
A $100 denim canine harness model struggled with low advert engagement. The issue was that rock patches have been on the again of the harness, so product photographs confirmed the canine from behind, hiding its face. Canine homeowners scrolling social media stored scrolling previous adverts with out cute canine faces seen.
So, Aditya Chavan and his crew used ChatGPT (with DALL-E) over 5 days to create AI-generated product photographs exhibiting each the harness design AND the canine’s face from flattering angles.
Subsequent, they scraped Amazon opinions of competing canine harnesses, used AI to investigate suggestions, recognized the highest 4 buyer priorities (good match, escape-proof, breathable, aesthetics), and rewrote advert copy round these.
There was an surprising discovery, too. New adverts with AI-generated photographs bought large engagement however low conversions. Utilizing Microsoft Readability to trace web site habits, they found why: the AI photographs appeared higher than the precise product. Prospects felt misled when the actual harness didn’t match the AI-generated promise.
As an alternative of abandoning AI photographs, the consumer redesigned the bodily product to match what AI had proven clients wished, creating the “Rocker 2.0” assortment based mostly on the AI-generated idea.
Because of this, every day orders and income each doubled, whereas the product redesign based mostly on AI insights efficiently matched buyer expectations.
Takeaway
You may analysis competitor opinions on Amazon or Google and use AI to identify widespread buyer frustrations and needs. Then, create photographs that talk on to these wants. Instruments like ChatGPT, Midjourney, or DALL-E allow you to shortly generate and check product concepts—no want for expensive photoshoots or giant stock upfront.
Create 10–20 picture variations in only a few days, run them as adverts, and use the outcomes to see which of them your viewers connects with most.
Advice
You should utilize a software like Ahrefs’ Key phrases Explorer to gauge precise consumer demand for particular product varieties and options. Enter seed key phrases like “canine harness” and discover the “Matching phrases” report. You should utilize the “Ecommerce” filter to slender down outcomes to precise merchandise. This tells you which ones options have actual demand, not simply which of them get talked about in opinions.
Allied International Advertising wanted to launch a marketing campaign for NERF Motion Xperience, the primary official NERF play heart within the US. NERF, recognized for its colourful foam dart blasters, had a facility nonetheless beneath building—no photographs or movies existed.
They used DALL-E to generate illustrations of teenagers with NERF blasters, nevertheless it wasn’t easy: they went by means of “too many iterations to rely,” closely edited outputs in Photoshop (isolating characters, adjusting colours), transformed every little thing to vectors in Illustrator, and filtered precise NERF merchandise to match the artwork type.
The end result: they met unattainable deadlines and created a full library of marketing campaign belongings because of AI.
Takeaway
AI can generate visuals when nothing exists to {photograph} but, providing you with unimaginable pace to market, however don’t anticipate plug-and-play. Work with a designer to scrub up outputs—that’s the place high quality occurs.
AGORA, a artistic company, created high-end marketing campaign visuals for jewellery, equipment, and CPG manufacturers by combining AI-generated surroundings with historically photographed actual merchandise.
As an alternative of a full photoshoot crew (artistic director, photographer, stylist, make-up artists, and so on.), they diminished it to: artistic director, immediate engineer/AI artist, and photographer/retoucher.
Attempting to AI-generate the precise product led to poor outcomes—the merchandise nonetheless want skilled pictures. Gimmicky all-in-one AI instruments don’t work for high-end tasks.
The successful system was to make use of AI for elaborate backgrounds (seashore scenes, studio setups, outside places), then composite professionally photographed actual merchandise into these AI scenes.
https://www.linkedin.com/posts/matteo-trapella-b104a913b_heres-how-we-harnessed-the-power-of-ai-to-activity-7313138354789801985-mTf0
Takeaway
AI can substitute costly location shoots, however not product pictures. Use AI to create background scenes. {Photograph} your product professionally, and use Photoshop to composite.
AGORA’s strategy is greatest for: jewellery, equipment, packaged items—any product needing premium presentation however versatile environments. Don’t use it for style on fashions or the place product/atmosphere interplay issues.
AI works for giant manufacturers which have lengthy stood for high quality content material, if used correctly. At Ahrefs, one of the vital revered website positioning blogs within the business, we use AI all through our content material workflow whereas preserving people firmly in management. Let me present you a number of examples.
Through the analysis section, instruments like NotebookLM assist shortly course of giant quantities of enter—equivalent to dozens of responses, interviews, or survey solutions—to floor the insights that matter.
AI assistants with “deep analysis” characteristic are nice for checking what’s already recognized a couple of subject and figuring out gaps price exploring.
I’ve additionally used ChatGPT and Claude to deal with extra technical duties equivalent to writing API code, combining datasets, or operating primary statistical checks.
We additionally use our personal Ahrefs’ AI Content material Helper to investigate top-ranking pages and maps which subjects want protection, eliminating the necessity to manually open dozens of tabs.
Not too long ago, we began making our personal vibe-coded inside instruments. For instance, I used Riff to make an app that inserts WordPress shortcodes and picture alt textual content (through Gemini API). And to assume I did all that manually…
Takeaway
Use AI to compress analysis time, automate formatting, and preserve consistency. At all times have people deal with artistic selections, fact-checking, and stopping generic AI voice. That is tips on how to use AI with out sacrificing content material high quality.
Gael Breton from Authority Hackers created an in depth system immediate defining his publication’s tone and construction, supplied actual examples of previous work, then fed the AI his uncooked, badly formatted notes and bullet factors—the unique, high-value insights.
The AI dealt with sharpening, writing, and even ran supplementary analysis utilizing Perplexity for sections like “Instrument of the Week.” He manually reviewed and edited all output.
This allowed them to chop publication manufacturing time in half (2 hours complete per week) whereas sustaining excessive reader engagement. Discover the division of labor: AI did formatting, sharpening, and researching simply obtainable details; the human supplied unique concepts, insights not documented on-line, and high quality management.
Takeaway
Use an AI assistant with an in depth system immediate that features your tone, type, and examples. Give AI your uncooked ideas and bullet factors, let it format and polish, however at all times evaluation and edit to stop a generic “AI voice”.
That is good augmentation, not substitute—you concentrate on producing proprietary insights whereas AI handles mechanical writing duties. Human experience creates worth, AI removes tedium.
Eric Siu from Single Grain used AI through dictation instruments to workshop video ideas. He fed the AI most popular thumbnail types with efficiency knowledge (like Mr. Beast’s view stats), requested for viral alternate options and full outlines, and let AI analyze patterns from profitable content material.
Brainstorming time dropped from hours to about one hour most, giving his design and enhancing groups clear, particular path based mostly on data-informed artistic selections.
However discover what didn’t change: Eric nonetheless made the ultimate artistic selections. AI dealt with sample evaluation, different era, and description construction; the human dealt with artistic path, ultimate selections, and model judgment.
Takeaway
Speed up artistic improvement with out sacrificing high quality—AI handles analytical work (sample recognition, different era) so people spend extra time on strategic artistic selections.
Advice
Eric was analyzing profitable content material for his movies—I’d take that additional. Use Ahrefs’ Content material Explorer to search out content material individuals hyperlink to most frequently in your area of interest. Search your subject and type by referring domains to see what naturally earns backlinks. Feed these prime performers to your AI and ask it to determine what makes them link-worthy—the info, format, or angle. Then apply these insights to your content material concepts.
Within the instance under, I’m searching for concepts on the subject of vibe coding in English, revealed this 12 months, excluding homepages for higher outcomes and sorted by variety of referring domains.
Life Media UK had lower than 24 hours to create a video marketing campaign for Puxton Park’s “Curly Tails & Cocktails” occasion, with zero current advertising and marketing belongings. Conventional filming was unattainable.
So, they used AI video era to create a promotional teaser, however made a essential strategic selection: they used stylized, whimsical animation as a substitute of making an attempt to look real looking.
By making it clearly AI-generated and artistic (not fake-realistic), they prevented the backlash manufacturers like Vogue confronted when making an attempt to go AI content material as actual. The video was “eye-catching, scroll-stopping” and efficiently promoted the occasion.
Takeaway
For last-minute campaigns with no footage, use AI video instruments (Runway, Pika) however go for “whimsical,” “animated,” or “illustrated” seems to be. Keep away from making an attempt to make it appear to be actual footage—that’s the place manufacturers get roasted. Suppose “cartoon”, not “documentary.”
When McDonald’s relaunched its Snack Wrap on July 10, 2025, Popeyes responded instantly. They commissioned PJ Ace’s AI studio Style.ai to create a diss observe video—and bought it carried out in three days.
The 64-second video encompasses a rapper dissing Ronald McDonald (“Meals be tastin’ humorous when a clown is in da kitchen”) in surreal celebration scenes with sloths, giraffes, and aliens consuming Popeyes wraps.
https://twitter.com/PJaccetturo/standing/1943321631272923229
To cover AI’s lip-sync limitations, they positioned a big pop filter over the rapper’s mouth in sales space scenes, turning a technical weak spot into an aesthetic selection that emphasised the vitality.
The marketing campaign was a viral success, racking up over 10 million views throughout platforms by mid-July and matching McDonald’s timing completely. Extra spectacular was the fee: roughly $200 in month-to-month AI software subscriptions versus the $50,000-150,000+ a conventional 3-day rush manufacturing would have required.
Takeaway
When timing issues greater than perfection, AI video instruments like Veo 3 and Runway supply a fast-track answer. Begin with an in depth script describing digital camera actions, lighting, and character actions, generate footage utilizing text-to-video AI, edit the clips collectively, and launch inside days.
This strategy shines for entertainment-first content material the place AI’s surreal, humorous aesthetics work in your favor. As PJ places it, ‘Entertain first, branding second’—comedic AI content material persistently outperforms makes an attempt at genuine or severe messaging.
Superside wanted a high-quality industrial with out time or finances for actors and a studio. As an alternative of asking AI to easily make a video, they constructed it the best way you’d assemble a collage, layer by layer.
They used Midjourney to generate character photographs, Runway to make these photographs transfer for 3-4 seconds, and their very own workers recorded strains with actual emotion on their telephones, then used ElevenLabs to “filter” their voices to sound like skilled actors. This stored human emotion however added skilled polish.
The end result: 40-60% sooner than conventional manufacturing, excessive conversions, and 0 bodily manufacturing prices.
Takeaway
Consider AI video like placing items collectively, not making every little thing from scratch. As an alternative of filming each scene your self, you’ll be able to construct a robust story utilizing totally different AI instruments and nonetheless hold the human components that matter most.
For instance, you’ll be able to create characters with Midjourney, animate them with Runway, and have actual individuals report the dialogue on their telephones. Then you’ll be able to evenly improve the voices with instruments like ElevenLabs so that they sound cleaner and extra skilled—with out dropping the actual emotion.
This combine works properly as a result of people deal with the emotions and efficiency, whereas AI takes care of the time-consuming visuals, animation, and polish.
Gael Breton used CapCut’s built-in AI to robotically flip lengthy movies into vertical clips. He then used automation to transcribe these clips with Gemini and had Claude write the descriptions.
He nonetheless threw out 3 out of each 4 clips, however the 25% that made it by means of averaged round 2,000 views per put up, and that content material wouldn’t have existed in any respect with out AI.
The method was sooner and cheaper than hiring full-time editors and writers, whereas human judgment was nonetheless key to deciding what was ok to publish.
Takeaway
You received’t like every little thing AI generates out of your prompts, however one thing like 25% and even 10% success price remains to be price it for content material that wouldn’t exist in any other case.
Very Eire, a significant on-line retailer, used AI to rewrite 1,471 product titles for his or her Google Purchasing campaigns. Working with Wolfgang Digital and Producthero AI, they ran similar campaigns side-by-side—one with handbook titles, one with AI-optimized titles.
AI efficiency began 3% worse in month 1 however ended 11% higher by month 4, with a 5% enchancment in return on advert spend. AI wanted time to be taught, however as soon as it did, it outperformed handbook optimization whereas saving 10 hours of work.
Takeaway
Use AI to rewrite titles in batches. Begin along with your prime 20-50 merchandise, give them your model tips and best-performing examples, then check AI titles in opposition to handbook ones in separate campaigns.
AI breaks the trade-off between personalization and effectivity. However be affected person; anticipate a studying curve of a number of months earlier than seeing full outcomes.
Past copywriting, AI may assist automate your advert workflows with instruments like n8n. For instance, my colleague Remus Marcu constructed n8n workflows that robotically test search phrases in Google Advertisements and flag whether or not they’re related, utilizing the marketing campaign’s key phrases, product descriptions, and previous detrimental examples.
Advice
If you wish to do this in your personal merchandise, I’d suggest utilizing Ahrefs’ Patches to rewrite web page parts at scale.
As an alternative of manually optimizing every product title one after the other, Patches permits you to replace tons of or hundreds of titles concurrently utilizing AI whereas sustaining your model tips. You may check AI-optimized titles in opposition to your present ones, measure efficiency, and roll again modifications if wanted—precisely what Very Eire did, however with far much less handbook work.
Barilla used AI software Buzzly to run a 15-day Meta promoting marketing campaign the place AI dealt with the testing and optimization that entrepreneurs usually do manually. As an alternative of making a number of advert variations and checking efficiency weekly, the AI generated a number of variations of advert copy tailor-made to totally different viewers segments, then robotically shifted the finances to better-performing adverts in real-time.
In simply 15 days, the AI-optimized marketing campaign drove 43,240 clicks at a 6.58% CTR—properly above the everyday 2-4%—whereas chopping cost-per-click by 81% (from $0.16 to $0.03) and decreasing administration time by 75%.
Takeaway
AI can automate the repetitive work of testing advert variations and adjusting budgets. As an alternative of manually creating 5-10 advert variations, operating them for every week, analyzing leads to a spreadsheet, then making modifications, you’ll be able to have AI generate dozens of variations, check them concurrently, and optimize in real-time.
Advice
Earlier than you let AI optimize something, be sure to perceive how your clients really speak and search. Begin by searching for the precise phrases they use. A software like Ahrefs’ Key phrases Explorer will help you discover high-intent key phrases in your area of interest and present how in style they are.
For instance, if “fast pasta recipes” will get 18k searches a month however “simple weeknight dinners” will get 23k, that’s a robust trace about which angle to strive first.
Moreover, you will discover extra angles to strive by exploring search time period variations within the Matching phrases report.
Coca-Cola constructed “Create Actual Magic,” a customized AI platform (combining GPT-4 and DALL-E) with OpenAI and Bain & Firm.
They gave digital artists worldwide entry to their most iconic model belongings—the contour bottle, Spencerian script brand, Santa Claus, and Polar Bear—and invited them to create unique paintings.
The marketing campaign ran as a contest the place artists may submit work for an opportunity to look on Instances Sq. and Piccadilly Circus billboards.
Coca-Cola chosen 30 creators for a three-day “Actual Magic Artistic Academy” at their Atlanta headquarters to co-create content material for merchandising and digital collectibles. All members have been credited.
Takeaway
AI can supercharge UGC contests by decreasing the barrier to entry—members don’t want design expertise, simply concepts. Conventional UGC campaigns require individuals to {photograph}, movie, or design one thing from scratch. AI-powered UGC lets anybody with a immediate create professional-looking content material that includes your model.
Closing ideas
In the event you’re getting carried away with AI hype, I believe these examples ought to carry issues again to actuality. None of those manufacturers fired their advertising and marketing groups or changed technique with a number of prompts. They nonetheless relied on individuals who understood their clients, made artistic calls, and caught AI’s errors.
However in case you’re skeptical—in case you assume AI is simply costly vaporware—have a look at what really occurred. Very Eire saved 10 hours month-to-month and boosted click-through charges 11%. Barilla slashed their cost-per-click by 81% and minimize marketing campaign administration time by 75%. Heads or Tails Pup doubled their every day gross sales. Popeyes bought 10 million views for $200 in AI subscriptions versus the $50,000+ a conventional rush video would price. Authority Hackers minimize publication manufacturing time in half.
Even when AI by no means improves past the place it’s right now, these time financial savings are actual—they usually add up quick over months and years.
Received questions or feedback? Let me know on LinkedIn.
