Noticias
Petit Comitech | The future of agentic commerce with Ander Orcasitas (Mastercard)
You want to buy a pair of sunglasses. You type into Google: sunglasses. Millions of results appear, some sponsored, along with a map showing dozens of stores where you can try them on. With so many options, it’s hard to choose.
In the near future (which is already starting), this won’t happen. You’ll talk to your trusted AI. In a matter of seconds, an agent will take you to a personalised website based on your preferences. And it will buy for you. “It will be an experience where you don’t have to think.”
Ander Orcasitas’ mind is already half a step ahead. He is VP of Dynamic Yield, Mastercard’s platform that enables brands to create real-time personalised digital experiences. Of course, powered by AI. It recommends products, adapts content and optimises interactions to boost loyalty and sales. This is what is known as “agentic commerce”.
The first edition of Petit Comitech, a series of exclusive meetings with tech leaders organised by Tech Barcelona and EAE Business School, resulted in a flood of ideas and our best attempt to distil them into this summary:
As Ronald McDonald used to say: “I’m Loving It”
The dream of many startups — being acquired — came true in 2019. Dynamic Yield was bought by McDonald’s. Suddenly, the chain’s menu is no longer a static board: it changes depending on the time, weather, how busy the restaurant is or the day’s trends. If it’s hot, it suggests a McFlurry. And you think: “I was just craving that.” What a coincidence… Everything leads to faster decisions, higher average tickets and customers leaving saying “I’m Lovin’ It”.
Less than a year after the acquisition, Dynamic Yield added more than 100 new brands across sectors ranging from e-commerce to finance and gaming, reaching over 350 clients in total and revealing a global need. In 2022, Mastercard saw the opportunity and acquired it.
Paradigm shift: AI enters the game
For years, the internet has been predictable. You searched, compared and chose. SEO and SEM structured that world — it was “deterministic”.
New AI models break that logic: “they are arbitrary, they never do the same thing twice”. Ander explains how systems decide for you, mix sources and prioritise without fully explaining why. We move from searching and finding to asking and trusting.
LLMs like ChatGPT have a double layer of power. On one hand, they control what the model shows and in what order. On the other, they receive user information in the form of conversation and search metadata. This exchange redefines the value of interaction.
As traffic coming from LLMs continues to grow (already accounting for 70% of searches in the Asian market) and competition for user attention increases, brands need to react.
Are we ready?
“The technology is already here.” The question is how ready we are to adopt it. For Ander, what matters is “retail readiness”: companies having the data, APIs and architectures in place to operate in a real agentic environment. He gives a two-year window for all websites to stop being static and start adapting in real time to each user. It’s either that or fall behind.
We could go into complex questions about who owns these technologies, what biases they carry, what dependencies they create and under what regulation they operate. But then the summary would stop being a summary.
We’ll soon share the full conversation between Ander Orcasitas and journalist Pipo Serrano.
