Coca Cola

Coca-Cola made headlines last holiday season when the company published its “Holidays are coming” ad, generated with AI.

The ad faced mixed opinions in the spectrum between “This kills human creativity” and “This is the future of storytelling.”

Despite the backlash, Coca-Cola continued its bullish intent on AI (which we think is important to navigate in this new world of intelligent workflows).

Anyway, two weeks ago, Coca-Cola published a more refined version of the 2024 ad.

This time, the ad was not only better, but the company also shared a BTS video of how only a team of five specialists refined over 70,000 videos to create the ad.

Coca-Cola is one of the earliest adopters of technology, and was listed in Forbes’ 50 AI early adopting enterprises list.

If you observe Coca-Cola as a customer, all you see is cool marketing campaigns.

But when you put your enterprise exec hat on, there is so much to learn about ops, innovation, and marketing with AI.

In today’s AI at the Top, we’ll unpack exactly that.

The enterprise has fundamentally shifted its marketing model.

Digital spend increased from less than 30% in 2019 to approximately 65% in 2024.

In fact, only two out of Coca-Cola’s 2000 marketing employees have AI in their title. This was a conscious decision to ensure AI is not a separate department, but a tech every employee uses.

So… where did Coca-Cola invest its marketing budget?

In 2023, Coca-Cola ran the “Create Real Magic” campaign in partnership with OpenAI and Bain & Company, inviting its customers to design how they see the brand.

It generated 120,000 pieces of UGC while the top submissions appeared on Times Square and Piccadilly Circus. This increased brand visibility, customer loyalty, and revenue.

In 2024, Coca-Cola created an AI Santa you can actually talk to in 26 languages across 43 markets. It was built in partnership with Microsoft.

The ROI is not publicly documented, but Coca-Cola’s VP and Head of Generative AI, Pratik Thakar, mentioned the ROI exceeded expectations with 1M+ customer engagements in 60 days.

Then in early 2025, Coca-Cola brought back the iconic “Share a Coke” campaign, with the popular names printed on the bottles. This is because Gen Z prefers authenticity and personalization.

The brand’s machine learning feeds on first-party and retail data allows it to decide which names and phrases to push out in the market.

Coca-Cola delivers 1.9 billion servings every day with nearly 225 independent bottling partners employing 700,000 people.

At this scale, it’s a suicide to not integrate AI and automations, and Coca-Cola knows this.

The company’s AI analyses historical sales, weather patterns, local events, and geolocation data to predict demand.

From the pilot program, CIO Neeraj Tolmare mentioned the forecast accuracy increased from 70 to 90%, increasing the sales. “We saw a 5-20% increase in MoM sales at roughly 1000 outlets where we tested this.”

Now, Coca-Cola is scaling demand prediction capabilities with its global bottlers.

In North American facilities, Coca-Cola uses robots for sorting and packing with automated vehicles for moving pallets and storage systems.

The brand has also optimised delivery routes based on real-time traffic, weather, and fuel costs. This reduced the last-mile fuel costs by 8%.

Not only delivery and sales, Coca-Cola also predicts maintenance and even shares insights with its retail customers.

AI sensors monitor production equipment and vending machines. Identifying issues before failures happen reduced the downtime by 20%

The brand has created an eB2B platform to connect with its 6.9 million retail customers.

AI generates push notifications and WhatsApp messages with advice on what stock to reduce waste and increase sales.

Coca-Cola uses real-time customer data to innovate its products.

The company launched Freestyle, a self-serve fountain system that lets customers mix and pour 100+ Coca-Cola beverages and flavour variants from a single machine.

50,000+ Freestyles are installed globally with 14 million drinks per day.

This data helps Coca-Cola understand customer preferences.

When AI identified customers frequently mix Cherry and Sprite, the enterprise launched Cherry Sprite as a standalone product. Orange Vanilla Coke and Sprite Lymonade were born from freestyle’s data as well.

In September 2023, Coca-Cola launched Y3000, the first flavour co-created with AI.

The company gathered human insights on what the future in the year 3000 tastes like (emotions, aspirations, colors, flavors, etc.) AI analysed the inputs, then the R&D developed the final formula based on AI’s flavour profiles.

YOU DECIDE

So far, the enterprise has adopted early and observed significant ROI on its AI investments:

Source: Klover.ai

The next big steps will focus on Coca-Cola’s partnerships with Microsoft and Adobe.

Unlike PepsiCo, which partnered with multiple enterprises, Coca-Cola’s partnership is only with Microsoft to create a unified ecosystem and global scaling. The partnership started in 2020 with $250 million and jumped to $1.1 billion for a 5-year second phase since 2024.

Post Microsoft, in May 2025, Coca-Cola announced Project Fizzion with Adobe to solve for brand consistency across 200+ brands.

Fizzion learns from how designers work inside Photoshop and Illustrator to speed up content production without losing brand integrity.

The partnerships, combined with Coca-Cola’s intent with AI, make the next three years an exciting time to observe the enterprise.

YOUR TURN