AI Is Moving into the Shopping Cart, Literally

Plus: It's starting to shape the purchase moment itself, not just the marketing around it.

Apr 07, 2026

Your daily signal on AI and CX — minus the hype.

📌 DCX Stat of the day: Carrefour Israel and A2Z Cust2Mate plan to deploy 4,000 smart shopping carts nationwide under a five-year agreement, with rollout starting in Q3 2026 across six flagship stores.

In this issue:

→ AI is now shaping the shelf
→ Where retail AI can go wrong fast
→ The human gap AI still has not closed
→ A simple test for AI-led journeys

🔎 Deep dive

Smart carts just made AI part of the store

For a while, consumer AI mostly lived around the journey. Search. Recommendations. Chat. Carrefour’s rollout is different. AI is moving into the basket itself.

Carrefour Israel and A2Z Cust2Mate signed a deal to deploy 4,000 smart carts with charging, software, implementation, training, and long-term support included. The deal also ties the rollout to retail media and data monetization rights. That matters because it moves AI from a support layer into the purchase moment itself.

The CX question changes fast when that happens. If the cart helps shoppers decide faster, the trip feels easier. If it starts pushing the wrong item, the wrong offer, or too much noise, the experience gets worse in the most valuable part of the journey. Once AI reaches the point of purchase, weak judgment starts to feel like pressure.

OPERATOR PLAYBOOK

Pressure-test one journey where AI can influence a sale

Pick one journey this week where AI is already shaping customer choice. Search to product page works. In-store assisted discovery works too.

Check four things:

Who owns the recommendation rules and priority logic?
Can you see where customers reject, ignore, or override AI prompts?
What happens when the recommendation is wrong, irrelevant, or misleading?
Can the customer recover without friction or staff rescue?

Ask your team: If our AI pushed the wrong choice today, who would know first and who would fix it?

Signal: Consumer AI gets expensive when nobody owns the correction loop.

📈 Market Reality Check

Emotional connection is a top customer value, and the weakest AI progress area

Only 16% of organizations report making meaningful progress with AI in building emotional connections with customers.

That is the sharpest signal in the ServiceNow study,The CX Shift: A Study of Customer Expectations in the AI Era. The company says the research covers more than 34,000 customers, service reps, and executives globally, and points to a widening gap between customer expectations and AI delivery.

For CX leaders, the implication is pretty clear. The near-term advantage may come less from replacing more human moments and more from using AI to clear low-value work so humans show up better in the moments that actually shape trust.

AI investment + weak emotional progress = a CX gap customers will notice

🧰 Tool Worth Knowing

Genie by Tote.ai

What it does: Loop Neighborhood Market is deploying Tote.ai’s Genie across its convenience store network. The assistant handles employee questions on customer service, POS operations, maintenance, payments, and more.

CX use case: This is internal AI with external consequences. If store staff get faster answers and rely less on managers for routine questions, customers usually get quicker help and more consistent service.

Why it matters: A lot of retail AI value may come from making the front line less hesitant, not from adding another flashy customer-facing layer.

Bottom line: The useful tools in consumer AI may be the ones customers barely notice because the store suddenly works better.

⚡ 90-Second CX Radar

Sezzle is bundling AI shopping and AI support into one consumer app

Sezzle said today that its app now includes an AI shopping assistant and 24/7 AI support as part of a broader push to become an all-in-one shopping, rewards, and financial education platform. The CX signal is that discovery, decision support, and service are starting to collapse into one consumer surface. That raises the trust bar because the same interface is now shaping both what people buy and how they get help.

Magicpin just launched an AI assistant for local retail

Economic Times reported today that Magicpin launched an AI assistant called Vera and committed $1 million to build out its AI stack for local commerce. The interesting part is not the assistant name. It is the direction of travel. Hyperlocal commerce players are trying to use AI to make smaller merchants feel easier to search, compare, and buy from. That could improve convenience, but it also puts more pressure on data quality and recommendation accuracy.

Alexa+ turns food ordering into one conversation

Amazon says Alexa+ can now place delivery orders through Grubhub and Uber Eats using a more natural back-and-forth flow, with browsing, customization, and order changes handled in one conversational window. The CX signal is pretty clear: task completion is moving away from command-based interfaces and toward adaptive, transaction-ready AI.

🧭 Your Move

The theme today is simple. AI is moving closer to where customers decide, buy, and ask for help.

Pick one journey where AI is already doing more than drafting or support search. Then map who owns the logic, who monitors bad outcomes, and how recovery works when the system gets it wrong. That is work a CX leader can do this week without waiting for budget season.