The Short Version First
In April 2026, Namly Design — a Nordic e-commerce brand selling personalized products — started running an AI agent platform for operational analysis.
The AI's first recommendation, after scanning the data, sounded counterintuitive: raise the free-shipping threshold.
The founder did it. Thirty days later: revenue up 23%, average order value up 16.1%, profit up 20.3%. And conversion rate? Down only 1.53%.
A stubborn fixation on maximizing conversion had quietly eaten into their profit for years.
What This Shop Does
Namly Design is a personalized-products e-commerce brand, selling across multiple country markets. The CEO is Peter Svendsen.
Before AI, they were already growing — 40% to 50% year-over-year. By most measures, doing fine.
But Peter felt something was off. Traffic was up, orders were up, but profit wasn't growing at the same pace as revenue. More money was coming in, but less of it was landing in his pocket than the top line suggested.
The Blind Spot the AI Caught
On April 14, 2026, they plugged into Ondigital's AI agent platform.
The AI's first analysis zeroed in on something seemingly trivial: free-shipping thresholds across markets.
It found a pattern — customers in different countries behaved differently on average order value and conversion rate. And Namly's existing threshold was tuned to maximize conversion: set low, so customers would easily place orders.
But the low-threshold trade-off was brutal: customers placed small orders, shipping ate a big share of each, and margins got squeezed. Conversion looked great. Profit quietly shrank.
The Counterintuitive Advice
The AI's recommendation was simple: raise the free-shipping threshold.
The logic:
- With a higher threshold, some customers who only shopped to hit free shipping drop off — conversion dips slightly
- The customers who stay place bigger orders — AOV rises
- Shipping's share of each order drops
- Overall margin improves
For a lot of e-commerce founders, raising the threshold and deliberately lowering conversion feels backwards. Who wouldn't want more people checking out?
Peter did it anyway.
The 30-Day Numbers
Comparing the 30 days before AI to the 30 days after:
- Revenue: +23%
- Average order value (AOV): +16.1%
- Profit: +20.3%
- Conversion rate (CVR): -1.53%
The headline isn't that conversion dropped 1.53% — it's that this tiny dip bought a 20% jump in profit.
And remember: historically this period was a down-trend for Namly. Instead of the usual 40 to 50% YoY growth, they hit 90% YoY.
What This Case Actually Shows
AI in e-commerce can do way more than generate product images and write copy.
The interesting thing here — the AI didn't generate new traffic, didn't pull any flashy trick. It just looked at the operational data and spotted a blind spot the founder couldn't see himself.
The team had stared at conversion for years, assuming higher was always better. The AI looked at the cross-market data and found that high conversion and making money weren't the same thing in this case.
But Some Cold Water
First, this is one month of short-term data. Whether it holds is another question. Raising the shipping threshold could affect customer loyalty and repeat purchase over the long run — 30 days won't show that.
Second, this is a case study published by Ondigital, the AI agent platform itself. The numbers are cherry-picked as a good story. In real deployments, AI gives bad recommendations too — the key is whether the human has the judgment to filter them.
Third, Namly isn't a huge operation. The one-insight-unlocks-20%-growth pattern may not replicate on larger, more complex platforms with messier data.
But this case hits a real issue: most small businesses don't lack data. They lack a pair of eyes that can spot the blind spots hiding in it. If AI can actually do that well, it's worth way more than a hundred generated product images.