Coca-Cola Uses AI to Tell Retailers What to Stock and How Much to Order

 

Coca-Cola is using artificial intelligence to help retailers decide which drinks to stock, how much to order, and when demand is likely to change.

In Malaysia, the beverage giant has expanded its Coke Buddy digital ordering platform with an AI-powered feature called Perfect Basket.

Instead of waiting for a retailer to decide which Coca-Cola products to reorder, Perfect Basket analyzes the business and generates personalized recommendations before the order is submitted.

The system can recommend both the products and quantities that may be appropriate for an individual store.

And it does not look only at what the retailer bought last time.

Perfect Basket uses Coca-Cola's Central Recommendation Engine, which analyzes previous orders, ordering frequency, seasonal demand, weather conditions, and purchasing trends among comparable businesses.

That means AI is starting to influence something extremely practical:

what actually appears on store shelves.

AI Is Helping Retailers Build the "Perfect Basket"

Coke Buddy already allows retailers to order Coca-Cola products through digital channels while accessing promotions, personalized suggestions, and order tracking.

In Malaysia, the platform currently supports approximately 39,000 retail outlets.

Perfect Basket adds another intelligence layer.

Suppose a convenience store normally orders Coca-Cola Original Taste, Sprite, and bottled water.

The AI may identify that:

hotter weather is approaching,

Sprite demand is increasing among similar stores,

a particular pack size performs better during certain months,

or the retailer may be underordering a product frequently purchased by comparable outlets.

Perfect Basket can then recommend a different product mix or higher quantity before the store submits its order.

Retailers are not forced to accept the AI recommendation.

They can review the suggestions, change quantities, remove products, or ignore the recommendation entirely.

The retailer still makes the final purchasing decision.

That human control matters because every recommendation involves real money.

Ordering too little can mean empty refrigerators and lost sales.

Ordering too much can leave cash trapped in inventory that sits unsold.

So an AI recommendation is not simply about predicting demand.

It is also helping retailers make working-capital decisions.

83% Followed the AI Recommendations During the Campaign

Coca-Cola recently revealed adoption figures following its Perfect Basket, Perfect Ride campaign in Malaysia.

The campaign ran from January to April 2026 and attracted more than 4,500 entries from over 4,000 retailers nationwide.

According to Coca-Cola Bottlers Malaysia-Singapore-Brunei, 83% of participating outlets adopted Perfect Basket recommendations during the campaign.

Participating outlets that followed the recommendations also recorded stronger sales revenue growth than comparable stores, according to the company.

But the number needs context.

The 83% adoption rate does not mean 83% of all 39,000 Malaysian outlets using Coke Buddy followed AI-generated orders.

It refers only to outlets participating in the campaign.

Coca-Cola also did not disclose exactly how much faster those retailers' revenues grew.

Nor did the Malaysian campaign release detailed figures covering:

forecast accuracy,

inventory reduction,

out-of-stock rates,

product wastage,

or logistics savings.

So the campaign provides evidence that retailers were willing to use the recommendations, but it does not yet provide enough public data to independently measure the system's full economic impact.

Coca-Cola Has Been Building AI Ordering Systems for Years

Malaysia is part of a much bigger Coca-Cola digital strategy.

Back in the first quarter of 2024, Coca-Cola said it and its bottling partners had connected nearly eight million customers worldwide to business-to-business digital platforms.

The company also said AI-powered suggested-order capabilities had already reached more than three million retail outlets in Latin America.

Retailers in India were also using AI-powered recommendations through Coke Buddy to place bulk orders.

This shows that Perfect Basket is not an isolated experiment.

Coca-Cola is gradually building AI into the infrastructure connecting the company, its bottlers, sales teams, and millions of retailers around the world.

Earlier Pilots Showed Higher Purchases of Recommended Products

Coca-Cola has previously reported encouraging results from similar AI systems.

During its second-quarter 2024 earnings call, the company said it was testing AI-generated personalized messages that recommended products based on retailers' previous orders and market data.

According to Coca-Cola, retailers receiving those messages were more than 30% more likely to purchase the recommended SKUs during initial pilots.

SKU means stock keeping unit, essentially a specific product and package combination.

For example:

Coca-Cola Original Taste in a certain bottle size,

Sprite in cans,

or Coca-Cola Zero Sugar in a specific multipack.

That 30% figure did not come from Perfect Basket in Malaysia, so the two results should not be treated as the same experiment.

But they show how Coca-Cola has been testing AI recommendation systems across different markets for several years.

Coca-Cola Is Combining Sales Data With Weather

Weather plays an important role in beverage demand.

People tend to buy different drinks depending on temperature, rainfall, location, events, and season.

Coca-Cola CIO Neeraj Tolmare told Fortune in 2025 that another AI demand-prediction project combined:

historical sales information,

weather patterns,

and geolocation data

to predict what retailers were likely to need.

Those predictions could then trigger messages telling store managers when they might need to replenish products such as Sprite or Diet Coke before inventory ran out.

Coca-Cola said a pilot across three countries generated sales approximately 7% to 8% higher than outlets that were not using the algorithm.

Again, that was a separate project from Malaysia's Perfect Basket program.

But the strategy is similar.

AI observes thousands of signals that might be difficult for an individual retailer to continuously analyze and converts them into a simple recommendation:

Order this product.

Order this amount.

Order it now.

AI Changes the Role of Coca-Cola's Sales Teams

This also changes what human salespeople do.

Traditionally, beverage sales representatives visit stores, check inventory, ask the retailer what needs replenishing, take orders, and suggest additional products.

Digital ordering reduces some of that routine work.

If AI can automatically prepare a likely order before the sales representative arrives, sales teams can spend more time helping retailers with:

product placement,

promotions,

store development,

new product introductions,

pricing,

and business growth.

Coca-Cola says its sales teams remain involved alongside Coke Buddy rather than being completely replaced by the digital system.

This is another example of how workplace AI is evolving.

AI does the repetitive analysis.

Humans spend more time on relationships, judgment, and higher-value decisions.

The Bigger Story Is Predictive Commerce

For years, e-commerce companies have used recommendation engines to predict what consumers might want to buy.

Now the same idea is moving deeper into business-to-business commerce.

Instead of Netflix recommending your next movie or an online shop recommending your next pair of shoes, enterprise AI can recommend:

what a supermarket should stock,

how many cases it should buy,

which products are likely to sell,

when inventory should be replenished,

and eventually how much stock should be positioned at each location.

That is predictive commerce.

Businesses are moving from reacting to demand after it happens toward trying to anticipate demand before it happens.

Coca-Cola's scale makes this particularly significant.

Its global system serves retailers across more than 200 countries and territories.

Every convenience store, supermarket, sari-sari-style neighborhood shop, restaurant, or food outlet generates potential signals about consumer behavior.

AI can turn those signals into increasingly precise recommendations.

What This Could Mean for Philippine Retailers

The Philippine implications are obvious.

The country has hundreds of thousands of sari-sari stores, convenience stores, groceries, restaurants, carinderias, cafés, and other small retailers.

Many still make inventory decisions based largely on:

experience,

personal observation,

supplier visits,

available cash,

and instinct.

And Filipino retailers are often very good at that.

The owner of a sari-sari store knows what the neighborhood buys.

But imagine combining that local knowledge with AI.

An intelligent ordering system could analyze:

previous sales,

payday cycles,

school schedules,

local fiestas,

weather forecasts,

holidays,

basketball games,

traffic patterns,

location,

seasonality,

and demand from similar stores nearby.

It could tell a retailer:

You normally sell more bottled water during this week's weather.

Demand for this pack size is increasing in similar stores.

You may run out of this product before the weekend.

Ordering another case today could prevent lost sales.

That could be especially valuable for small retailers where cash flow is tight.

A large supermarket can afford some inventory mistakes.

A sari-sari store operating with limited capital often cannot.

If AI can help a small entrepreneur buy closer to the right quantity, it could reduce waste, prevent stockouts, and help protect working capital.

But there is also an important warning.

AI recommendations should serve the retailer, not simply maximize how much inventory suppliers can sell.

A responsible system should be transparent, explain recommendations where possible, protect business data, and allow retailers to reject suggestions easily.

That is why Coca-Cola's current approach, where the retailer retains final control, is important.

Because the owner still knows something the algorithm may not.

Maybe there is a barangay fiesta tomorrow.

Maybe classes were suspended.

Maybe a competitor opened across the street.

Maybe the refrigerator is broken.

AI can analyze data.

But the retailer understands the reality on the ground.

The most powerful system will combine both.

AI Is Quietly Entering Everyday Business

When people talk about artificial intelligence, the conversation often revolves around ChatGPT, robots, AI videos, or job displacement.

But some of the biggest economic changes may look much less dramatic.

AI is quietly deciding:

which package should be delivered,

which shipping route should be used,

which customer needs attention,

which product may sell tomorrow,

and now,

how many bottles of Coca-Cola a retailer might want to order.

That is where AI becomes truly consequential.

Not because it generates impressive answers on a screen.

But because it starts influencing millions of small decisions inside the real economy.

For Coca-Cola, Perfect Basket is an ordering tool.

For the larger AI industry, it represents something much bigger.

Artificial intelligence is moving from recommending what consumers should buy to recommending what businesses should sell.