Artificial intelligence has long been associated with software, data centers, and the companies building large language models. But another wave of AI adoption is quietly transforming an industry far removed from Silicon Valley: consumer packaged goods. The world’s largest makers of everyday products — the businesses behind the bottles of shampoo, the bars of soap, and the packages of cookies that line grocery store aisles — are now using AI to design their products, optimize ingredients, and run the marketing campaigns that sell them.
This shift represents a significant expansion of AI’s role in industrial research and development. Instead of being confined to code and algorithms, AI is being embedded into the physical chemistry of shampoos, the taste profiles of biscuits, and the creative direction of advertising. For companies like Procter & Gamble (P&G), Mondelez International, and Unilever, AI has become a central tool for compressing timelines, cutting costs, and personalizing consumer experiences at scale.
From lab bench to algorithm
Historically, developing a new shampoo formulation or a new cookie recipe was a slow, iterative process. Chemists and food scientists would mix batches, test them physically, and rely on trial and error to refine the product. A single formulation could require months of work before it was ready for consumer testing. AI changes that equation entirely.
P&G offers a clear example. The company reported using AI to screen tens of thousands of peptides — short chains of amino acids — while developing a formula for a Pantene product. The screening relied on an internal database of more than 8,500 existing formulations, allowing the AI to predict how a given mixture would feel on skin or hair before any physical trial was conducted. This computational filtering drastically reduces the number of candidates that need to be physically tested, shortening the path to consumer trials.
The benefit is not novelty for its own sake; it is speed and efficiency. Steps that once required rounds of physical testing can now be narrowed down computationally, pushing the most promising candidates toward human trials faster than ever before. P&G’s approach mirrors a broader trend across the industry: treating product development as a search problem over known ingredient combinations.
Mondelez, the snacking giant behind brands such as Oreo, Cadbury, and Chips Ahoy!, describes a similar transformation on the food side. The company says it has deployed an AI product-development tool capable of generating dozens of new formulations at once. According to Mondelez, this software allows developers to work between two and five times faster than conventional methods. Instead of testing one or two new recipes at a time, researchers can simultaneously evaluate a wide array of variations, tweaking ingredients such as flour ratios, sweetener levels, and fat content to meet specific taste and texture targets.
The speed gains are not limited to formulation. AI is also being used to simulate how a product will behave during manufacturing — how dough will handle on a conveyor belt, how a shampoo will flow through a bottling line — reducing the need for physical pilot runs. This end-to-end acceleration is reshaping the R&D departments of consumer goods companies worldwide.
Marketing gets a generative overhaul
The same generative systems that drive product innovation are now being pointed at marketing. For decades, advertising campaigns required significant time and creative resources: developing concepts, shooting video, writing copy, and producing variations for different channels. AI now offers a way to produce personalized images, text, and video at a pace that traditional studios cannot match.
Unilever has leaned hardest into this application. Its Dove brand recently ran a cookie-scented body-care line in collaboration with Crumbl, the popular cookie chain. AI was involved across the entire effort — from product direction to the selection of influencers and even the creative assets themselves. The company reported that the campaign drew billions of impressions and brought a large share of new buyers to the brand. Whatever one makes of a cookie-scented soap, the mechanics are instructive: a single AI-assisted pipeline running from formulation to feed.
Generative AI allows marketers to produce thousands of variations of an advertisement — different backgrounds, different text overlays, different voiceovers — in a fraction of the time it would take a human team. This enables hyper-targeting: a consumer in Tokyo might see an ad with local landmarks and language, while someone in New York sees a version tailored to their preferences. The same technology can also select influencers based on audience data, predicting which personalities will generate the highest engagement for a given product.
The AI-driven marketing push is not limited to Unilever. At the Cannes Lions International Festival of Creativity, OpenAI pitched AI-made advertisements to agency executives, showcasing how large language models and image generators can produce campaign concepts from a brief. The message is clear: AI is becoming a core part of the advertising toolkit, and consumer goods companies are early adopters.
The business case: compression and cost
What ties these examples together is compression. In consumer goods, the traditional cost of experimentation is measured in months of lab work and test batches. The traditional cost of a campaign is measured in agency hours and production days. AI attacks both.
Reformulation becomes a search problem over known ingredients. A company like P&G can query its internal database for peptides that are likely to deliver a desired sensory property, then validate only the top candidates in the lab. Content creation becomes something generated and varied on demand, allowing brands to test multiple messages quickly and iterate based on real-time performance data.
This compression translates directly into financial benefits. Faster product development means more frequent product launches, keeping shelves fresh and maintaining consumer interest. Cheaper marketing production means brands can afford to target smaller segments with tailored messages, potentially increasing conversion rates.
The trend is also being fueled by a broader reallocation of enterprise budgets. Large companies in every sector are investing in AI agents and tooling. Tencent, for example, has developed enterprise agents that automate customer service and back-office tasks. In consumer goods, similar investments are flowing into R&D and marketing technology. The packaged-goods sector, long seen as a laggard in digital innovation, is now catching up.
Cautions and limitations
Despite the enthusiasm, the claims deserve caution. Most of the specific figures cited come from the companies themselves, and consumer giants have every reason to present their AI programs as further advanced than they truly are. Product development still ends with human tasting panels and dermatological testing. A formula that an algorithm predicts will work is not the same as one that a shopper buys twice.
Industry researchers have also flagged that AI-generated marketing often drifts toward the generic. Because generative models are trained on broad data, their outputs can lack the brand-specific character that makes a campaign memorable. A witty tagline or a distinctive visual style can be hard to replicate algorithmically. Without careful human oversight, campaigns can feel bland or formulaic.
There are also ethical considerations. AI systems that predict consumer preferences rely on vast amounts of personal data, raising privacy concerns. The use of AI to generate realistic images and videos also opens the door to misleading advertising if not managed carefully. Consumer goods companies will need to invest in governance frameworks to ensure that their AI-driven processes remain transparent and compliant.
Another limitation is the quality of internal data. AI models are only as good as the datasets they are trained on. If a company’s historical formulation database has gaps or biases, the AI may produce suboptimal or unsafe recommendations. For example, an algorithm trained primarily on successful products may fail to predict issues with novel ingredient combinations that have never been tried before.
Industry context: AI in product development
The use of AI in product development is not entirely new. Pharmaceutical companies have been using computational drug discovery for years, screening millions of molecules for potential drug candidates. Consumer goods companies are now following a similar playbook, adapting techniques from drug discovery to the design of shampoo, soap, and snacks.
This convergence is enabled by advances in machine learning, particularly deep learning and generative models. Tools like AlphaFold have demonstrated that AI can predict protein structures with high accuracy, a capability that directly translates to peptide screening for cosmetic ingredients. Similarly, generative models for chemistry, such as those used by companies like Iktos and SRI International, are being repurposed for flavor and fragrance design.
Beyond product formulation, AI is also being used to optimize packaging, supply chains, and retail placement. A shampoo bottle’s shape, color, and label design can all be tested virtually before any physical prototype is made. This reduces waste and allows for rapid iteration.
The consumer goods sector is also experimenting with AI-driven demand forecasting. By analyzing social media trends, weather data, and historical sales, companies can predict which scents or flavors will be popular in a given season or region, enabling them to adjust production accordingly.
Looking ahead: faster shelves, more variants
For shoppers, the visible result of this AI transformation will be mundane but pervasive: more product variants, faster refreshes, and scents and textures that appear and disappear from shelves more quickly than they used to. A single brand might now release a limited-edition flavor every quarter instead of every year. A shampoo line might offer a dozen different formulations tailored to specific hair types or local preferences.
The machinery behind the shelf is changing even where the products look the same. A bottle of shampoo is, increasingly, the output of a search — a search across chemistry, consumer data, and manufacturing constraints. That search is being conducted by algorithms rather than lab technicians, and it runs continuously rather than in discrete trials.
The direction is consistent across firms that rarely agree on much. P&G, Mondelez, and Unilever are competitors in many categories, but they are converging on the same AI-enabled approach to innovation. The packaged-goods sector is not sitting out the AI revolution; it is embedding it into the very products that line the aisles of every supermarket in the world.