This week, Wallapop, a Spanish platform, announced a strategic partnership with Swiss AI company Albatross to deploy real-time AI discovery across its platform.
With Albatross, Wallapop becomes one of the first C2C marketplaces globally to move beyond static, history-based recommendation systems toward in-session, adaptive discovery that understands user intent as it unfolds.
In practice, this means discovery on Wallapop becomes more pro-active and responsive. For instance, when a user explores a second-hand sofa and then starts browsing rugs or floor lamps, the system understands the shift from evaluating a single item to furnishing a living space. Instead of continuing to surface more sofas, as traditional recommenders do, discovery adapts in real time to reflect the user’s evolving intent, highlighting relevant complementary items across categories, including products the user didn’t know existed.
As the session continues, the experience keeps evolving: a user who then clicks on a vintage coffee table may start seeing lighting, paintings, décor, or storage pieces that match the same style, even if they never looked for them explicitly. This allows users to uncover highly relevant listings that feel like genuine finds, rather than variations of what they’ve already seen. Crucially, this allows discovery to keep pace as interests change (even within the same session) and helps the right listings surface at the right moment.
The announcement follows Albatross’ recent $12.25 million fundraise to deploy its real-time perception layer; a technology designed to understand what users want in the moment. Forbes recently described this approach as the “second pillar of AI.” While generative AI has transformed how content is created, Albatross focuses on how content is discovered in environments where scale and choice overwhelm traditional personalization.
In a four-week A/B test on 10% of Wallapop’s traffic, measuring performance across the entire homepage experience, the platform yielded:
- 118.9% increase in user engagement;
- 104.8% increase in favorites and interactions; and
- 46.9% increase in purchase intentions.
Importantly, the system surfaced previously unseen items to qualified buyers, demonstrating its ability to unlock latent supply rather than simply amplify already popular listings. These gains have remained remarkably stable – and continued to improve – over an extended four-month production period, underscoring the robustness of the approach beyond short-term experimentation.
For Wallapop, the shift goes beyond buyer relevance. The partnership introduces algorithmic seller discovery as a core capability, dynamically matching live buyer intent with relevant, shippable listings and addressing one of recommerce’s hardest challenges: ensuring quality supply doesn’t get lost in volume.
“At Wallapop, we are moving toward a system that understands what users want in real time, helping buyers find the right items faster while giving sellers more effective visibility for their listings,” said Rob Cassedy, CEO of Wallapop. “Our collaboration with Albatross represents another step forward in our mission to empower people to participate in a more conscious consumption model that creates economic opportunities for people.”
Albatross’ technology continuously interprets live user interactions. Every action within a session updates the system’s understanding of user intent and the catalog, allowing discovery to adapt instantly to changing behavior, supply dynamics, and context. This approach is particularly helpful for second-hand marketplaces, where millions of unique items with inconsistent metadata make traditional popularity- or similarity-based methods ineffective.
For Albatross, Wallapop represents a demanding real-world environment for AI discovery. With a constantly evolving catalog of one-of-a-kind items and high listing velocity, Wallapop provides an ideal setting to validate and scale in-session perception models under real marketplace constraints.
“Together with Wallapop, we’re delivering an online discovery experience comparable to the best in-store journey; one where users uncover products that feel like true hidden gems, tailored just for them, in real time” said Dr. Kevin Kahn, CEO of Albatross AI. “While large language models make sense of words, Albatross’ perception model makes sense of the sequence and context of user actions as they happen. This partnership demonstrates a radical shift in discovery, search, and personalization, which have seen little innovation for more than a decade.”
