:quality(75))
The "Sneaks" session at Adobe Summit 2023 is where Adobe shows what is still experimental. Two previews are relevant if you sell anything where discovery is the hard part: True Colours and Fast Filtered Commerce.
True Colours
True Colours uses machine learning to recommend products based on a customer's colour preferences. Plenty of people shop by colour; not many storefronts let them. True Colours works by:
analysing customer interaction patterns with coloured products
learning individual preferences from browsing and purchase history
recommending products matching preferred colours across categories
adapting in real time as preferences change
If you sell fashion, home or beauty — categories where colour is often the deciding factor — that is a direct route to larger baskets. Early testing showed engagement increases of 20-30% for merchants using colour-driven recommendations.
Fast Filtered Commerce
Fast Filtered Commerce rethinks product filtering. Conventional filtering hands the customer every facet at once: size, colour, price, brand. Fast Filtered Commerce decides which ones matter to that customer by:
predicting which filters matter most based on browsing behaviour
surfacing the relevant facets first
learning from click patterns which attributes carry weight
In pilot implementations it reduced time-to-purchase by 15-25%. For a merchant selling across diverse categories, filter presentation adapts on its own — dimensions and materials lead for furniture, size and colour for fashion.
The technology underneath
Both run on Adobe Sensei, so there is no custom model for you to build or train. Sensei learns from aggregated, anonymised data across the Adobe Commerce ecosystem, which means a smaller merchant gets intelligence trained on a volume of transactions they could never generate alone.
Privacy
Both were designed privacy-first: data collection is transparent, customers can opt out, and no data leaves customer infrastructure unless you explicitly enable it. For European merchants under GDPR, that is built in rather than bolted on.
What this means for you
These were previews, with availability expected later in 2023 — nothing to implement yet. If you want early access, that conversation goes through your Adobe account team. If you sell in fashion, beauty, home or across multiple categories, it is worth a line on the roadmap now so you are not starting from scratch when it lands.