Tesco’s proposed Czech and Slovak sale deserves a closer look: the €1.7bn headline includes leases, while expected cash proceeds are smaller. Ahold Delhaize’s prize is a larger supermarket network, with integration and regulatory approval still ahead. We also examine what AI shopping statistics actually measure, and why consumers’ willingness to accept help is a different question from giving software permission to buy.
News
1️⃣ Tesco agrees to sell its Czech and Slovak businesses to Ahold Delhaize.
2️⃣ Clarity detects checkout events, which do not establish independent AI purchases.
3️⃣ Merchant interviews find few orders from personal AI shopping agents.
Insights
4️⃣ Shoppers welcome AI assistance but mostly keep control of purchasing.
5️⃣ Europe’s largest webshops offer sharply different delivery choices.
6️⃣ Four European grocery markets split on trips and spending per visit.
LET’S ENTER THE MAZE!
1️⃣ News
The Maze: British grocer Tesco has agreed to sell its Czech and Slovak businesses to Dutch supermarket group Ahold Delhaize. The €1.7bn valuation includes leases; the deal still needs regulatory approval.
Source: Ahold Delhaize. Inside a Slovak Tesco store included in the proposed acquisition.
The Details:
Cash is smaller than the headline. Tesco expects €1.2bn before tax and costs. It plans to return most net proceeds to shareholders through extra share buybacks after completion.
Ahold wants local scale. The target includes more than 360 directly operated stores and over 100 franchises. Ahold plans to combine the Czech operation with its Albert supermarket chain, aligning ranges and loyalty offers.
The handover is still ahead. Ahold targets Q4 2027, subject to clearances. Hungary stays with Tesco, which says the sale sharpens its focus on the UK and Ireland.
MarketMaze Take: For retailers elsewhere, this is a capital-allocation choice: Tesco wants cash returns and focus; Ahold wants buying scale. The same stores can serve different owners’ priorities.
Selling does not prove failure. The two countries generated most of Tesco Central Europe’s £115m adjusted operating profit. That regional total includes Hungary, so it cannot establish the sold businesses’ exact profitability.
Scale must become useful. Aligned stores and loyalty offers could strengthen Ahold’s proposition. But expected savings are not achieved savings, and the announcement does not establish cheaper groceries or better supplier terms.
2️⃣ News
The Maze: Microsoft Clarity, a free website analytics tool, records more checkout activity in visits involving AI agents. That does not establish that software completed more purchases.
Source: Microsoft Clarity. Original study table showing agent-involving session counts, percentages and the denominator definition.
The Details:
People remain in the picture. Clarity’s study reports Checkout events in 25.21% of agent-involving visits, versus 14.65% of human visits. It cannot separate human takeovers from agent actions.
The base is unexplained. The study covers over 5.2 million agent-involving visits. Its checkout count and percentage imply roughly 1.22 million. The publication does not reconcile them or disclose the human visit count.
Labels are not receipts. Clarity’s event tracking groups clicks and page visits into named actions. Site administrators can customize the rules. The study does not establish completed autonomous orders.
MarketMaze Take: The commercial opportunity is better diagnosis of website friction. A larger checkout-event percentage alone cannot show that agents generate more sales.
Count the finish. Merchants anywhere should distinguish an interaction label from a confirmed order before using these figures in sales forecasts. No country breakdown supports a global uplift claim.
Keep the useful signal. Agent-assisted visits can still reveal where customers need help. The missing purchase evidence limits the business case; it does not prove those visits are worthless.
3️⃣ News
The Maze: Personal AI agents can shop for people, yet many businesses report few resulting orders. WIRED’s interviews across roughly 50 companies show why registrations cannot substitute for sales evidence.
Source: Meta. A frame from Meta’s September launch demonstration shows Muse recommending a stroller and displaying ordering messages; it does not measure actual merchant sales.
The Details:
Orders remain small at named merchants. DoorDash described orders from Muse and other agents as numbering in the hundreds. Hilton called bookings insignificant. These self-reports are not a representative adoption survey.
Meta counts something different. It says over 10 million people in the US and Canada set up Muse. More than a quarter of people using it engaged in shopping. That does not disclose total completed orders.
Checkout still matters. Restaurant chain Chipotle reported agents struggling with loyalty discounts and after-hours orders. It was fixing website code. An agent finding a meal is not the same as successfully buying it.
MarketMaze Take: The early opportunity is making authorized purchases work, while testing whether they bring sales. The evidence does not justify treating agent shopping as either inevitable disruption or a failure.
Measure the last step. Merchants elsewhere can track completed orders and repair observed failures. US and Canadian account growth cannot establish worldwide demand.
Keep the objection visible. Some retail partners had not yet shared post-launch results. This interview sample may miss concentrated activity; it cannot settle the channel’s eventual size.
4️⃣ Insight
The Maze: AI shopping agents can find products, follow instructions or buy for you. Across 16 countries, consumers welcome help much more readily than independent purchasing.
In Accenture's survey, 85% welcome collaboration and 74% would delegate specific tasks. Both keep the shopper involved.
Permission narrows: 32% would let AI choose while they pay; only 9% allow independent buying. These measure willingness, not usage.
BrightEdge's client data show 43% monthly growth in agent purchases, yet just 3% of its ecommerce clients allow agent transactions. Neither figure describes all online commerce.
Why it matters: Make products easy for agents to understand, then separate selection from payment approval. Helping shoppers does not require taking away their final say.
5️⃣ Insight
The Maze: The delivery menu that works at home may miss local expectations. Seven European markets show big differences in what the largest webshops offer at checkout.
Locker delivery appeared at 91% of the largest Estonian webshops and 9% of Belgian ones in October 2026: an 82-percentage-point gap in retailer availability, not shopper demand.
Collection methods differ too: Denmark had parcel shops at 79% and lockers at 13%. Parcel shops are staffed pickup points; lockers are self-service compartments.
Estonia also had home delivery at 97%, so offers overlap. Tembi's checkout research measures retailer menus; sample counts and the exact large-shop cutoff are undisclosed, and unclassified options make shares minimums.
Why it matters: Benchmark similar local retailers before adding delivery partners. Check price and service promises too: an available option does not prove profitable demand.
6️⃣ Insight
The Maze: More grocery trips can mean less money per visit. Europanel’s four-country comparison shows why retailers should separate shopping frequency from basket spend.
France’s trips rose 4.3% and spend per trip fell 2.7%; Germany moved +3.5% and -2.4%. These changes compare the first half of 2026 with 2025.
Croatia and Poland both made 3.9% fewer trips, while spend per trip rose 8.0% and 3.8%. Each visit therefore carried more money, with different increases in the two markets.
Europanel’s figures cover take-home everyday goods. Bigger baskets mean higher spending, not necessarily more items; prices and product choices may also contribute.
Why it matters: Check visits and spend per visit before copying a neighbouring market’s retail plan. More buying occasions and more money per basket call for different tests.
🧰 Resources
Useful stuff, handpicked by the team every day
🤖 AI prompts: 14 AI Prompts for International Expansion & Localization or 👉 All here
📖 Glossary: Shipping & Delivery, Decoded: 17 Terms You Actually Need to Know or 👉 All here
📅 Events: Digital 1to1 Italy Spring 2027 or 👉 All here
🔓 Get the full resource library with MarketMaze+
🗞️ Quick hits
Everything else you should know
🛒 Retail economics
U.S. online grocery grows as store sales fall: online food-and-beverage sales rose 15.5% in the four weeks ending October 4; physical-channel sales fell 2.8%. Total sales still slipped, so digital growth did not mean a growing market.
UK charity shops are earning less despite growing secondhand demand: Oxfam’s reported retail profit fell 23%, while the industry disputes blaming resale apps over costs and donation quality.
📱 Platform businesses
Apple cuts Taiwan’s standard App Store commission to 26%, leaving developers 74 rather than 70 per 100 units of eligible app purchases. Existing 15% tiers stay unchanged, so the gain excludes sales already on reduced rates.
Affle agreed to buy Perion for US$420 million. The Israeli adtech company can seek rival offers until December 11; the buyer plans a contract review as it shifts toward charging for customer actions.
🚚 Delivery
UPS says chip-enabled labels now cover 3 million parcels daily at 315 customer sites, giving earlier pickup confirmation. Its 6 million daily-parcel figure is a 2027 target, not current volume.
Delivery app downloads fell 21% across DoorDash, Uber Eats, Instacart and Grubhub in the US and Canada in 2026 to date, Sensor Tower estimates. Paid sources supplied 25.9% of installs; downloads do not measure orders.
💳 Payments
OPay postponed its planned US stock listing after citing a regulatory inquiry. The Nigerian payments platform says transfers and payments continue; that assurance is not an independent service check.
THAT’S IT FOR TODAY!
You’re the reason our team spends hundreds of hours every week researching and writing this email.
What do you think of this issue?
See you next time in the maze!
MarketMaze team
