Global Digital Marketing & Retail by Alex 89
Inspiration from across the world for retail enthusiasts, e-commerce professionals, marketing lovers and technology fans. Welcome back! I summarized some great links again, I stumbled upon this week.
Goodmorning! Today lot’s of AI news, so go start reading!
🌪️How AI agents reshape
Let’s start with a substack of a fellow substacker on AI agents. Regular readers of my newsletter know I follow this topic closely. Checkout for example my summary of the salesforce agentic world tour .
“Every abundance creates a new scarcity. The task for builders is to figure out which resources AI makes abundant, which are rendered scarce, and where an edge can be formed. So far, for those who have found an edge, the business results are incredible.
AI Agents Challenge SDLC: Unlike traditional software, AI agents use flexible, goal-based models, creating unpredictable results.
Upgrade Instability: Traditional updates are smooth; LLM updates can disrupt agents, requiring re-training.
New Paradigm: Moving from deterministic, affordable software to adaptive, costly AI agents poses new reliability challenges.
The web created jobs like UX designers and PMs.
AI agents will do the same.
AI doesn’t replace people, AI + human replaces people, raising the quality bar.
A few examples:
Agent Engineer: Building, deploying, and optimizing agents.
Agent Architect: Crafting effective and engaging agent conversations.
AI Workflow Designer: Designing workflows that use agents to automate tasks.
Ethics Officer: Ensuring agents behave ethically and comply with regulations.
Agent Integration Developer: Connecting agents to various tools to make them more useful.
💳Visa, Mastercard step into AI agent payments
As we just talked about agents, an important part in commerce agents is the payment. Visa and Mastercard now build the payment rails for the (near) future. When giants like this do that, things are speeding up, but it is equally as important in my opinion to study and read about open source alternatives, and to realise that these “agentic”models can make mistakes – sometimes referred to as 'hallucinations' – leading to unintended and potentially expensive orders for example.
It’s important to understand this risk, whatever payment rails you choose.
AI-Ready Cards: Replaces card details with tokenized digital credentials, enhancing security for consumers and simplifying payment processes for developers. They confirm that a consumer’s chosen agent is allowed to act on a consumer’s behalf and bring identity verification to AI commerce.
AI-Powered Personalization: The consumer is in control. Consumers share basic Visa spend and purchase insights with their consent to improve agent performance and personalize shopping recommendations.
Simple and Secure AI Payments: Allows consumers to easily set spending limits and conditions, providing clear guidelines for agent transactions. Commerce signals are shared in real-time with Visa, enabling Visa to effect transaction controls and help to manage disputes.
🦾All in podcast: 90% of AI work is work that we do not do today
And now we are at it, a (slightly) different perspective from the All-in boys, I sliced a snippet of the complete podcast:
..The vast majority of the use cases are, are is the ability to finally review the contracts that we never got around to reviewing. It's the ability to finally automate an invoice process that we never did. It's the ability to go in and just create marketing campaigns in every language that we never got around to. And so, so I think that'll probably be actually like 90% of the usage of AI in the future will be things that if we look back and we snap the line right now, we said, this is what knowledge work is today. 90% of AI usage will be things that we don't do today.
Open AI is planning to charge between two and 20k a month for different levels of AI agencies would be basically cron jobs.
Chamath expresses concern about the transition from deterministic software (where the outcome of an action is predictable) to probabilistic software like Large Language Models (LLMs) that can "hallucinate" or generate incorrect information. He emphasizes that while traditional software follows a defined sequence of steps (do A, then B, then C), LLMs can produce unexpected errors….
…This shift makes quality assurance (QA), including unit and integration testing, critically important – even more so than before.
…Flexport has a large network of truck drivers using their app, but they don't have enough loads to keep all drivers actively checking for matches. Manually calling each driver to inform them about relevant loads was too expensive, even using call centers. AI calling has now made it cost-effective to contact thousands of drivers daily about potential loads, activating drivers who might have otherwise missed opportunities, representing entirely new work….
Complete pod here
Manus agentic use cases mentioned in the pod
📖Reddit Q1 Report: dive in those metrics
Reddit’s Q1 report is out. Not only it’s important to know for marketers (as it can have huge impact on SEO) but also I think it is interesting to understand their metrics. I also love their openness and the way how they manage to get most people to a logged in status. Worth a read!
For seekers, Reddit’s open nature is essential—it allows our content to surface across the open web and be easily found in search. We remain one of the last major platforms that doesn’t require you to sign in to learn something because we believe that by giving everyone access to knowledge, we are helping fulfill the purpose of the internet. This openness broadens visibility, drives awareness, and brings us new users— but it also means that some of our traffic from external sources is variable. “
Some metrics they use:
Daily Active Unique Visitors
Daily Active Uniques: U.S. Versus International
Logged-in Daily Active Uniques
Logged-Out Daily Active Uniques (DAUq)
Average Revenue Per Unique (ARPU) also increased
Details: https://s203.q4cdn.com/380862485/files/doc_financials/2025/q1/Q1-25-Earnings-Press-Release.pdf
🔍Great great slide deck on SEO and user experience
Go check it out! Another really good one, not only for pure SEO but also it gives advice for branding and user experience. It includes tips for AI tools to rapidly improve your product content!
Complete deck via Speakerdeck: https://speakerdeck.com/aleyda/the-state-of-ecommerce-seo-how-to-win-in-todays-products-serps-number-seoweek
Or via my dropbox: https://www.dropbox.com/scl/fi/04dy7yyxi5w7pl2rti0l4/stateofecommerceseo-seoweek-aleyda___ToDeliver.pdf?rlkey=6ssjod0a3p69rte8sxig57mc2&st=2b3oanvc&dl=0
🇨🇳 Meanwhile in China: everything becomes a billboard
The coolest billboards can be found in Asia! But what now comes from China is impressive. I think it is not completely new, I saw it even in Asia during holidays, I think, but this gained traction in the West recently: a flexible, transparent digital film that can turn windows, signs, and even random walls into glowing, moving ads. Think of all the use cases in stores, public transport or buildings. Curious when we get to see it here in Europe in and on stores and public transport.
Also from China..
Also from China from Hangzhou, we see Chinese and French guests speaking in their own languages, with AR glasses from a company called “Rokid”. These glasses incorporate advanced language models (LLMs) and use AR projection technology to overlay translations and other information, with a battery life of about 4 hours after a quick charge. This seems so great, so much better understanding in the world due to this type of communication!
Details: https://global.rokid.com
👴Wise words on product management: one to save!
I remember when I did PM, it’s indeed difficult to distinghuish yourself from backlog management or to “own” the backlog. People expect it from you, but that’s not the essence. Don’t go get to process orientated, stay focused on customer’s problems. Make sure the backlog reflects the customer’s problems and prioritize it also that way.
Source: https://x.com/nurijanian
🌐The (coming) translation boom?
I love localisation. It’s great to localise products and to feel you add value to users. This graph below indicates that there is so much potential:
The average number of languages supported by the 150 global websites studied in this report is 34. As you can see here, we’ve largely plateaued in languages — despite the fact that there remains demand around the world for many hundreds of languages.
Whatever we call it, AI promises the next wave of language growth. So now is the time to develop a world-ready website architecture and a content strategy that is unbound by language. These may be uncertain times but what remains certain is the need for websites and apps that speak the world’s many languages.
Details: https://globalbydesign.com/2025/04/29/deglobalization-ai-the-calm-before-the-storm
That’s it for this edition, thank you for reading, liking and subscribing
Curious for more immediatly?
Check out my archive of all previous newsletters: