AI retention flow = $15.2K in one quarter

The booking bot that made chatbot skeptics sweat

Master of Code built Aveda a Facebook Messenger booking chatbot that obliterated every expectation. We're talking about a 767% increase in weekly bookings and a 378% jump in lifetime users—all in seven weeks.

Here's what's got the industry buzzing: this wasn't some fancy GPT-4 integration. It was advanced natural-language processing that actually understood what customers meant instead of just matching keywords.

The contrast? While Aveda automated their biggest bottleneck—appointment scheduling complexity—most eCommerce operators are still building FAQ bots that make customers want to throw their phones.

Why it actually worked: The bot handled the full booking workflow from discovery to confirmation, eliminating multi-step friction that kills conversions. Think about how many times you've abandoned a booking because the process was too annoying.

What to do now: Find your biggest conversion killer and build AI to eliminate that specific pain point. Budget $15,000-30,000 for proper NLP implementation instead of wasting $3,000 on keyword-matching toys.

Source

1 billion tasks a day and counting

Daily API invocations with Alibaba International's AI system hit 1 billion as of July 2025. Up from 1 million in 2023. A thousand-fold increase in two years.

Their Marco AI now generates nearly 40% of all search-optimized product descriptions across Alibaba International's platforms. That number's climbing toward 50% faster than anyone expected.

But here's what's really wild: Marco doesn't just handle one thing. It crushes over 60 specific tasks across marketing, marketplace compliance, and customer service. In 40+ languages. While many of us are still copy-pasting product descriptions like it's 2019.

The three AI agents changing everything:

• Intelligent Refund Agent: Cuts refund costs by 15% through market analysis

• HS Code Agent: Automates customs classification with 23% better accuracy

• Merchant Recruitment Agent: Nearly doubles email lead conversion rates

Here's what should concern Amazon sellers: API calls from platforms outside Alibaba's ecosystem increased 23-fold as everyone scrambles to license Marco's capabilities. This isn't just optimization—it's infrastructure other platforms can't compete without.

Your move: Start automating the obvious stuff. Jasper AI or Copy.ai for product descriptions runs $50-100/month versus hours of manual writing. Goal: eliminate 40% of content creation tasks by December.

Source

The death of the "no results found" page

Miva launched Vexture AI search that understands what shoppers mean, not just what they type. The AI classifies queries into Specific, Broad, and Sparse categories, then triggers different response logic for each. Translation: no more lost sales from typos, slang, or vague descriptions like "brake line thingy."

What makes this breakthrough: Vexture combines local private AI embeddings from merchants' product catalogs with real store data. When someone searches "eco-friendly water bottle," they find reusable bottles even if those exact words aren't in the product title.

The technical advantage everyone's missing: this runs natively inside Miva with zero third-party dependencies. No API delays, no external processing, no monthly fees to AI companies that might disappear tomorrow.

Customer expectations vs. reality: Traditional search forces customers to think like machines. Vexture handles contextual understanding—"running shoes for trails" surfaces relevant products without exact matches. It learns continuously from your catalog and customer behavior without manual tuning.

Every "no results" page costs you $50-200 in abandoned sessions. Miva merchants can preview AI classifications and go live with one click.

What to do now: Request Vexture beta access if you're on Miva. Everyone else should implement Searchspring or Klevu AI search ($300-800/month) to capture those conversions.

Source

From Excel hell to $15.2K extra revenue

2(X)IST (the underwear brand) was manually running RFM analysis—Recency, Frequency, Monetary value—in spreadsheets. Hours of work to figure out which customers were likely to churn, which were loyal, which needed nurturing.

Then they automated the whole thing with AI. Result: $15.2K in incremental revenue in one quarter through targeted retention flows.

Winners vs. laggards: Instead of static Excel analysis, their RFM segments now update automatically in real-time. The system tracks when customers move between segments and triggers personalized retention campaigns with discounts calibrated to individual customer lifetime value.

The contrast: instead of sending the same "20% off" email to everyone, they now send different offers to first-time buyers versus VIP customers who've spent $2,000+ over two years.

The scalability factor: What used to take hours of manual work now happens automatically across their entire customer base. The AI spots micro-segments and behavioral triggers that humans miss in spreadsheet analysis.

What to do now: Deploy Klaviyo's RFM automation ($20-100/month) or Segments.ai ($300-500/month). Stop treating new customers and repeat VIPs the same way.

Source

What the smart money is actually doing

B2B companies are investing 11-25% of their eCommerce budgets in generative AI, according to McKinsey's latest research. One in four dollars going toward AI initiatives.

And it's working. One marketplace they studied grew GMV by more than 50% from September 2023 to September 2024 using shared AI infrastructure across markets with local customizations.

Reality check: B2B decision makers typically use ten or more distinct channels during purchase journeys. AI helps manage this chaos by understanding customers across all touchpoints instead of treating each interaction like it's happening in a vacuum.

The talent problem nobody talks about: Companies struggle with "unclear understanding of what tech talent they need and what 'good' looks like." DHL's response? Building a European Innovation Center focused on AI, robotics, and automation.

The cookie connection: Companies are scrambling to implement AI customer understanding before Google kills third-party cookies in 2025. This isn't optional—it's survival.

What to do now: If you're spending less than 10% of your eCommerce budget on AI, you're falling behind B2B leaders. Focus on customer preparation tools first.Source

What customers expect vs. what they get

62% of millennials prefer visual search over typing to find products online. But here's the twist: only 34% of shoppers will let AI assistants make purchases on their behalf.

Translation? Customers want AI to help them discover and research, but they want control over the final transaction. Makes sense when you think about it.

The infrastructure race: Adobe found 65% of senior eCommerce executives believe AI and predictive analytics are key to growth strategies. And 60% of retailers are increasing AI infrastructure investment in the next 18 months.

What's possible vs. what's happening: AI-enabled sites see 47% faster purchase completion compared to traditional eCommerce. But 40% of shoppers get frustrated when there's no human assistance available during AI interactions.

The opportunity: Voice search complements visual discovery by handling reorders and basic questions. The combination creates more complete customer journeys than either channel alone.

What to do now: Test visual search for your top categories—optimize product images with clear, high-resolution photos. Implement voice-friendly descriptions that work conversationally.

Source

THIS WEEK'S ACTION

1. Find your biggest conversion bottleneck and build AI to solve that specific problem.

2. Test AI search because every zero-result page costs $50-200 in lost sales.

3. Deploy RFM automation before your competitors do.

4. Audit your AI budget - some say that if you're below 10% of eCommerce spend, you're behind.
   

There’s no shortage of AI newsletters out there — but I couldn’t find one made for small and mid-sized eCom operators. So I built the kind of thing I’d want to read. Hope you get something useful out of it.

Thanks for reading.
Catch you next week!

Nicholas Hoddevik
CEO, apparel brand Kattnakken
Editor, BeyondTheCart