The global conversational commerce market is projected to reach $32.6 billion by 2035, growing at 14.8% CAGR. McKinsey estimates that generative AI could unlock $240–390 billion in economic value for retailers. UK grocery retailers risk losing £500 million per week to AI-driven alternatives if they fail to adapt (Algolia, 2025). OpenAI has already launched Instant Checkout in ChatGPT with Target and Instacart — the market is moving toward conversational commerce. And yet, most grocery AI initiatives never make it past the pilot stage. Grocery retail is ready for AI. But which AI, and for what?
According to Statista, 77% of grocery shopping interactions now begin on a smartphone. OpenAI has already launched Instant Checkout in ChatGPT with Target and Instacart, signalling that the world’s largest AI companies see conversational commerce as the future of retail. The question is no longer whether AI will transform grocery — it’s whether your business will be the one leading the change or playing catch-up.
The problem: traditional sales channels are no longer enough
A supermarket’s e-commerce site converts 1-3% of visitors. The app requires download, registration, and navigation. The call centre costs £3-5 per order. Meanwhile, WhatsApp has a 98% open rate and over 3 billion monthly active users. From Tesco to Sainsbury’s, from Walmart to Albertsons, every major grocery retailer is investing in digital transformation. The question is not whether your customers use WhatsApp — they already do. The question is: are you using it to sell?
Consider the numbers: a typical grocery app has a 30-day retention rate below 25%. The average customer visits a supermarket’s website 2-3 times before abandoning the cart. Phone ordering costs £3-5 per transaction in staff time alone. Meanwhile, WhatsApp messages are read within 3 minutes on average, with response rates that dwarf every other channel. For grocery retailers like Tesco and Sainsbury’s investing millions in digital transformation, the channel with the highest ROI is already in every customer’s pocket.
What is conversational commerce in grocery
Conversational commerce in grocery retail means letting customers order as they would speak to a person. “Do my weekly shop.” “Add something for a Sunday roast.” A voice note from the kitchen. A photo of a product from the pantry. The AI understands, remembers, suggests — and sells. This is not a chatbot with buttons. It is a system that manages thousands of products, weight variants, dynamic pricing, delivery zones, loyalty discounts, and family preferences in a natural conversation on WhatsApp.
The distinction matters because the complexity of grocery is unlike any other retail vertical. A fashion chatbot handles hundreds of SKUs with simple attributes. A grocery AI must orchestrate thousands of products with weight variants, daily price changes, perishable inventory, recipe-driven ordering, dietary restrictions, and regional preferences — all in real time, all in natural language. This is why generic AI assistants fail in grocery: they are not built for this level of operational complexity.
Why WhatsApp and not an app or a website
WhatsApp is already installed on every customer’s phone. No download required, no login, no training. The customer writes as they would to a person — and the AI responds like a shop assistant who knows every customer by name. With a 98% open rate versus 20% for email, WhatsApp is the channel with the best cost-to-conversion ratio in grocery retail.
In markets like India, Brazil, and Southern Europe, WhatsApp is effectively the internet. In the UK and US, adoption is growing rapidly, particularly among the 25-45 demographic that drives grocery spending. Walmart’s partnership with delivery platforms and Albertsons’ digital initiatives both point to messaging as the next frontier. The key advantage of WhatsApp over proprietary apps is zero friction: no download, no account creation, no learning curve. The customer writes “do my weekly shop” and the AI does the rest.
What it takes to move from pilot to production
An AI system for grocery retail is not a generic chatbot trained on FAQs. It is an engine that must manage: catalogues with thousands of SKUs and variants, prices that change daily, promotions by customer segment, delivery zones with time slots and tiered costs, minimum order values, free delivery thresholds, heavy-item surcharges, seasonality, and the memory of every individual customer — what they buy, how often, for how many people, with which dietary preferences. 90% of AI projects fail because they cannot handle this complexity. The difference between a prototype and a production system is entirely here.
Most AI pilots fail because they test with 50 products in controlled conditions, then collapse when faced with a real catalogue of 10,000+ SKUs. A production system must handle a customer sending a voice note in a noisy kitchen, requesting “the usual plus something for the kids’ packed lunch,” while the AI simultaneously checks inventory, applies a loyalty discount, suggests a substitute for an out-of-stock item, and calculates same-day delivery availability. This is not prompt engineering — it is systems engineering at scale.
How to evaluate an AI platform for your supermarket
We created a framework of 18 questions that every grocery executive should ask any AI technology vendor.
The framework covers nine critical dimensions: language understanding, multi-modal input, customer intelligence, product intelligence, business logic, conversation management, security, technical architecture, and production readiness. Each question is designed to reveal whether a vendor has solved a real operational problem or is merely demonstrating a controlled scenario. The difference between the two determines whether the system will generate revenue or generate support tickets.
GroceryAI: in production in grocery retail since 2026
GroceryAI is the only platform that combines all of these capabilities into a single operational system. It has been in production in grocery retail since January 2026, processing real orders with real payments and real deliveries. It is not a prototype, not a research project — it is the AI sales channel of a retailer serving hundreds of customers every week.
The platform orchestrates 9,000+ products across 4 synchronised warehouses, enforces 100+ business rules automatically, and processes orders in any language through text, voice, and photo input. Customer memory tracks the 200 most-purchased products per customer, enabling one-message reordering of an entire weekly shop. Early deployment data shows a 15-25% increase in average order value compared to traditional e-commerce, driven by contextual suggestions and persistent personalisation.