• by Laurie Sullivan , May 29, 2026

Shopsense AI will soon announce a Shoppable Intelligence Model (SIM), the company’s next-generation intelligence multimodal artificial intelligence (AI) model, which powers real-time contextual commerce.

“The industry has done a great job understanding audiences,” Bryan Quinn, president and co-founder at Shopsense AI. “If you have a dog, you may shop for pet food and then look around the internet for travel content. Even when looking at the travel content, it may make sense to serve the consumer travel content, although travel may not be directly related to the dog.”

"SIM,” the core underlying technology, has quietly launched with its first public partner, People Inc.

Quinn said the more information the technology can provide a media buyer, the easier it becomes for them to make major decisions like optimizing in real time for programmatic that was not previously possible in the past, Quinn said.

"The industry has been measuring progress in AI by how well models recognize content,” said Tory Brangham, People chief ecommerce officer. "The real benchmark is whether they can monetize it. SIM raises that standard by tying intelligence directly to commerce outcomes, such as intent and conversion.”

Other clients include Fox Tubi, Paramount, Disney, and Univision, as well as others the company will announce soon.

Shopsense tested SIM against publicly available models such as OpenAI's CLIP and Google's SigLIP2, and found it delivers between 25% to 50% higher retrieval accuracy across benchmarks for product discovery.

For retailers and marketers using SIM, lift arrives automatically across every storefront and in-content commerce units. Lift results in actual real revenue lift with no additional effort.

For publishers and site visitors, the model delivers more relevant recommendations, higher audience engagement, and stronger commerce revenue, with no changes to existing integrations.

On Fashion-200K and FashionGen, two public fashion datasets used by the research community to track progress in retrieval AI, SIM outperformed each open-source baseline across every retrieval modality such as image-to-image, image-to-text, text-to-image, and text-to-text.

AI product recommendations rely on subtle visual cues such as the cut of a sleeve, the texture of a fabric and the curve of a heel as well as precise vocabulary used to describe them.

A retrieval accuracy improvement of 10 percentage points on a benchmark translates into recommendations that drive shopper behavior changes.

For every 100 customer searches, 10 additional shoppers see exactly what they were looking for as the first result, company data shows, and that translates into click-through conversions and time to purchase.

At the scale of a live publisher or retailer network, this compounds across click-through, conversion, and time-to-purchase.

In Shopsense engagement data, this has resulted in a 10% improvement in model precision, which produces a 24.5% improvement in shopper click-through rate for native retail-media activations.

SIM has been optimized for fashion and apparel, accessories like jewelry and footwear, furniture and core commerce categories.