AI in marketing examples: 10 brands using AI

The best AI in marketing examples include: Coca-Cola for speedy localization, Heinz for audience engagement, Adidas for personalization, and Avis for customer support. come from brands that published a result, not just a launch. This article walks through 10 of them, from generative creative at Coca-Cola to predictive ad audiences at Pegasus.

Marketing and sales are the functions where companies most often report AI improving revenue. According to McKinsey’s 2026 survey, 37% of respondents said their companies saw revenue gains from using AI in marketing and sales.

Below, we share 10 AI marketing examples from real brands with tips you can copy yourself.

AI in marketing examples at a glance

This table helps you find the AI example closest to your own channel and goal. It compares 10 brands by AI application, channel, objective, and verified outcome, from email copy at Farfetch to WhatsApp support at Avis.

10 real-world examples of AI in marketing

Brands are using AI to create campaign assets, personalize product recommendations, improve advertising, and handle customer interactions. The 10 examples below show how this works in practice, and each brand has shared figures related to its results.

We’ve organized the list by use case instead of industry:

• Generative AI and content creation
• AI personalization and predictive recommendation
• Conversational and AI shopping experiences
• Predictive advertising and campaign optimization

Generative AI and content creation examples

Coca-Cola used AI to speed up localized design production, Heinz boosted engagement with AI-generated visuals, Unigloves created 250 targeted product images, and Farfetch improved email performance with AI-generated copy.

Generative AI can help teams produce visuals, adapt creative assets, and write email copy faster. These four examples show how brands use it, including ways to get customers involved in campaigns.

1. Coca-Cola scaled creative production with generative AI

Objective: Coca-Cola wanted to produce localized campaign variations faster, while keeping a consistent brand identity.

AI application: It worked with Adobe to build Project Fizzion. The system learns how Coca-Cola’s designers make creative decisions and uses those rules to create on-brand variations for different markets.

Outcome: Coca-Cola says Fizzion helps creative teams produce content up to 10 times faster. When the company announced it in May 2025, the system was still in the pilot stage.

Speed wasn’t the only goal, though. Coca-Cola wanted designers to stay in control of the creative work.

The system also goes beyond using a brand guide as a reference. It applies rules to design elements, so logos and typefaces can adapt to different layouts, languages, and formats.

What marketers can learn from Coca-Cola: Teach AI how your team makes design decisions, rather than feeding it a brand guide and hoping for the best. Faster production helps, but someone still needs to set the rules and review the results.

2. Heinz turned AI-generated visuals into a participatory campaign

Objective: Heinz wanted to show that people instinctively associate ketchup with its bottle. It also wanted to turn that idea into something people would share.

AI application: Working with Heinz, its agency Rethink tested an image generator with prompts like “ketchup in space.” The results kept looking like Heinz bottles. Heinz then asked the public to send in their own prompts and turned the images into a campaign. Fans who sent prompts got special-edition bottles with AI-made labels.

Outcome: According to Rethink’s award entry, the 2023 campaign earned more than 850 million impressions. The entry values those impressions at ~2,500% more than the media investment. The engagement is reported to be 38% above previous Heinz campaigns.

Those figures tell us about reported reach and engagement, rather than sales. They’re also agency-submitted results, not audited data.

What marketers can learn from Heinz: Give people a reason to interact with your campaign. Heinz started with an AI experiment, then let customers turn it into something bigger by adding their own ideas.

3. Unigloves created targeted product images with generative AI

Objective: Unigloves wanted to market its Derma Shield product to people in different professions. That meant creating product visuals for various settings.

AI application: The company worked with Superside to generate images for five distinct scenarios. AI gave the team a way to create these visuals affordably.

Outcome: According to Superside’s report, the team produced 250 unique images and estimated a 57% saving in design hours. That percentage reflects the team’s estimate rather than a measured result.

Each scenario called for a different working environment, even though the product stayed the same. Traditional photography would have required separate shoots, locations, and talent for each setting.

What marketers can learn from Unigloves: Think about how many different ways customers use your product. If you need visuals for several audiences, use AI to create them without repeating the same production process.

4. Farfetch used AI to optimize email copy

Objective: Farfetch wanted to improve its email subject lines and campaign copy to get more people to engage with its emails.

AI application: Farfetch used Phrasee to test different subject lines and email copy. The tool learns how a brand communicates and suggests copy that fits its style. Farfetch used it for both promotional campaigns and automated emails, such as browse abandonment, basket abandonment, and wish list reminders.

Outcome: Chain Store Age reported a 31.1% increase in open rates and a 37.9% increase in click rates for triggered and lifecycle campaigns. The figures come from a case study published in 2022.

Broadcast emails saw smaller gains, with open rates rising by 7.4% and click rates by 25.1%. That difference suggests AI copy had a greater impact on triggered emails than on broadcast campaigns in Farfetch’s results.

What marketers can learn from Farfetch: Try AI-generated copy on triggered emails first. These emails go out in response to customer actions, giving you regular opportunities to test different versions and compare results.

AI personalization and predictive recommendation examples

Personalization models read behavior, purchase, context, and preference data. They use it to predict what a customer is likely to want next. These three brands apply that to products and experiences.

5. Adidas used AI to personalize product recommendations

Objective: Adidas wanted to make its homepage more relevant to individual shoppers and help them find products they might want to buy.

AI application: The brand used machine learning to rank products for each visitor. The system considers browsing and purchase signals, then uses them to decide which products to recommend.

Outcome: A case study from its technology provider reports a 13% increase in homepage conversion rate over one month.

That figure comes with a few caveats. The provider doesn’t share the original conversion rate, and the case study is undated. The test also lasted just one month, so the result doesn’t establish whether the improvement lasted.

What marketers can learn from Adidas: Personalization doesn’t have to start with a complicated store-wide project. Test recommendations on one important page first, then use your conversion data to decide what to do next.

6. Zalando used AI to recommend products and find the right fit

Objective: Zalando wanted to help shoppers find products and sizes that suit their needs.

AI application: Zalando built its own AI models using fashion data to match shoppers with relevant products. It calls this approach personal matchmaking. The company also used Size & Fit AI to recommend sizes based on real body measurements from over one million customers.

Outcome: Zalando’s March 2026 annual results report a 13% increase in items added to bags. The company also reports that Size & Fit AI reduces size-related returns by more than 8%.

The figures offer a useful signal, but the release doesn’t explain the test methods or baselines. Add-to-bag activity also doesn’t show how many shoppers completed their purchases.

What marketers can learn from Zalando: Give shoppers help beyond product suggestions. Size guidance may help reduce avoidable returns, so track both sales and returns when measuring personalization.

7. Spotify turned customer data into personalized experiences

Objective: Spotify wanted to use listening data to make music discovery more personal and give users experiences worth sharing.

AI application: Discover Weekly picks songs based on each listener’s habits, including what they save and skip. The AI DJ adds spoken commentary between tracks. Wrapped gives users a personal look back at their listening year.

Outcome: At its 2026 investor day, Spotify reported nearly 20% more interaction with DJ messages. It also logged more than 620 million Wrapped shares in 2025.

The AI DJ had reached 94 million Premium users across more than 75 markets by May 2026. That’s reach rather than a performance lift, so read the two figures separately.

What marketers can learn from Spotify: Give customers a personal experience they want to show their friends. When people share it, your brand reaches new audiences through their own posts.

Conversational AI shopping examples

Conversational AI helps customers feel more confident in their choices by answering questions that might stop a customer from purchasing the product or service. 

8. Avis handled customer inquiries with conversational AI

Objective: Avis wanted its team to spend less time on routine questions and more time helping customers who needed it.

AI application: The company used WhatsApp to answer common questions about rentals and bookings. Customers could chat with the AI assistant first and get help from a person when needed.

Outcome: Insider One’s case study reports that the assistant handled more than 70% of inquiries and delivered 39% cost savings over a year.

What marketers can learn from Avis: Put support on a channel customers already use.  Let AI handle the questions that don’t need a human.

9. Amazon used Rufus to help shoppers find what they need

Objective: Amazon wanted to help shoppers find the right products without leaving the app to look for answers elsewhere.

AI application: Rufus is Amazon’s AI shopping assistant. People can ask questions in their own words instead of searching with exact product terms. Rufus can also shop millions of products from other online stores through Buy For Me. Auto Buy takes things a step further by purchasing a product when it hits a price the shopper has set.

Outcome: Amazon says more than 300 million customers used Rufus in 2025. The company estimates it helped bring in nearly $12 billion in incremental annualized sales. 

That $12 billion figure comes from Amazon’s own estimate. The company hasn’t explained how it separates Rufus-driven sales from other purchases.

What marketers can learn from Amazon: Chats and clicks don’t tell you whether an AI assistant makes money. Track the orders it helps bring in, then compare them with a baseline to see if sales increased.

Predictive advertising and campaign optimization examples

Predictive AI helps advertisers focus on people who are more likely to buy. Pegasus Airlines used this approach to find audiences with a stronger chance of booking.

10. Pegasus improved advertising ROAS with predictive AI audiences

Objective: Pegasus Airlines wanted to spend its ad budget on people more likely to book a flight.

AI application: The airline used predictive models to estimate which users were most likely to buy. It then focused its ads on those audiences instead of treating every user the same way.

Outcome: Insider One’s case study reports a 17% increase in return on ad spend over two months.

What marketers can learn from Pegasus: AI needs real customer data to spot patterns in buying behavior. Early results can look promising, so give campaigns more time before drawing firm conclusions.

How to use AI for ecommerce email marketing?

Ecommerce marketers can use AI in email marketing to create targeted audience segments, personalize product recommendations, and write email copy. That’s where ecommerce marketing platforms can help. Omnisend Agent AI features cover five common tasks: segmentation, product recommendations, content creation, campaign automation, and customer insights. 

Here’s what you can expect:

• Segmentation: Tell the Omnisend Agent which customers you want to reach, and it will create a segment with filters matching contacts that you can review and adjust before using.
• Product recommendations: Omnisend Agent uses customers’ browsing and purchase history to choose relevant products. Each person sees different recommendations in the same email.
• Content assistance: The Agent drafts email copy in your brand’s tone and suggests subject lines. You can edit the drafts and pick the suggestions that fit your campaign.
• Campaign automation: The Agent reviews your automated campaigns and recommends relevant fixes in timing, conditions, content, and other elements to improve performance. The suggestions can vary from adding an exit condition to creating additional workflows for underused segments. 
• Omnisend MCP: Omnisend connects to your favorite AI tools like Claude or ChatGPT so you can ask about campaign performance, identify improvement opportunities, and generate content in one chat window.

How to get started with AI in marketing

To get started with AI in marketing, choose a measurable goal, check that you have the necessary data, test AI on a small scale, review its outputs, and track results against your baseline. 

• Choose one measurable problem: Decide what you want to improve. It could be open rates, conversions, or the cost of handling customer inquiries.
• Confirm you have the data: Product recommendations need purchase history, while predictive audiences need conversion data. So, make sure you have enough before investing.
• Test on a limited workflow: Run AI on one page or workflow first. This ensures a single change explains the result.
• Keep a human in the loop: Review every AI output before it reaches a customer. Keep a log of what you changed.
• Track the result against a baseline: Always record where you started. Otherwise, you’ll have no way to prove what changed.

For practical setup steps, Omnisend’s help center guide to getting started with AI walks through turning each feature on. Here’s a video that explains Omnisend’s AI features.

Examples of AI in marketing: FAQs