The Depict H2 2026 public roadmap
In H2 2026, Depict unites Visual Merchandising, Search, and Recommendations into a single Shopify app, deploys its first AI agent, and rebuilds analytics into an active optimization layer. Here is the quarter-by-quarter view of what's coming.
In H2 2026, Depict unites Visual Merchandising, Search, and Recommendations into a single Shopify app. Moreover, we continue building for operational efficiency: deploying our first AI agent and giving you total control over how your brand appears, while our AI takes over the tedious work of keeping discovery optimized.
One Depict on Shopify
In Q3, we're bringing the full power of Depict to Shopify. Visual Merchandising, Search, and Recommendations will be available together in a single unified Depict app. One install, one interface, one place to run product discovery across your store. Each product can be adopted on its own or together, so you can start where the need is sharpest and expand as it makes sense.
Next-generation search
Starting in Q3, merchants gain deeper control over search results: explicitly exclude or promote specific products, shape results the way you shape a collection page. We're introducing dynamic query suggestions, advanced autocomplete, and the ability to index entire collections directly within search results. For brands that lead with imagery, product titles and prices can optionally be hidden from the search grid for a cleaner, more visual experience.
In Q4, we launch Multimodal Search, letting shoppers combine image and text to find exactly what they're looking for, alongside AI Overviews: nuanced, comprehensive search summaries delivered directly on your storefront.
Personalized recommendations and dynamic PDPs
We're bridging automated intelligence and manual curation, and we're deliberate about which is which.
Merchants get expanded configuration for manual recommendation overrides, plus session-based Dynamic Recommenders that adapt what each shopper sees based on what they've browsed and engaged with during their current visit, within rules you configure.
By Q4, this evolves into Looks: mini-collections mapped across your catalog using relational graphs, embeddable anywhere from search results to product detail pages. Alongside Looks, our new Personalization Engine enables Dynamic PDPs and PLPs: layouts that adapt to the shopper based on rules you define. Surface sale items and bundles for budget-driven sessions; lead with immersive lifestyle imagery for visually-driven ones. You set the personas and the rules; the engine executes them in real time.
The Depict Agent
We're rolling out a Depict Agent that helps you keep surfaces optimized while keeping you in full control.
Fixes it handles on its own. Pure discovery hygiene, the kind of work that's unambiguous but mind-numbing at catalog scale, so it usually doesn't get done. A red umbrella showing up in results for "black shoes." "Sneakers" and "trainers" being treated as unrelated terms when they should return the same products (Auto-Synonyms). These have a right answer and no brand judgment involved, so the agent simply fixes them.
Changes it proposes, you approve. Anything that touches how your brand shows up. If a product has high impressions but poor click-through, the agent might propose a Visual Swap, switching its thumbnail to a lifestyle shot, but the change waits for your sign-off. Your storefront's look stays a human decision; the agent's job is to surface the opportunity and do the legwork.
Actionable analytics
Data is only as valuable as the actions it inspires. We're rebuilding our analytics suite from passive charts into an active optimization layer.
Merchants will track click-through rate, no-results rate, and search refinement rate (how often shoppers have to adjust their query after the first attempt), with localized dashboards across markets. We'll also show search conversion rate alongside site conversion rate, so you can see how discovery performance sits in the context of your store's overall numbers.
Rather than just displaying data, the system processes it into directive insights through a One-Click Optimize framework:
Insight: Product X has high search impressions but zero clicks. Suggested action: Update its primary thumbnail image. Option: Let the Depict Agent handle fixes like this going forward.
Connect Depict to your AI chat with MCP
Q3 brings a big update as we enable you to connect Depict to your favorite AI agents and execute visual merchandising operations (generating collections, auditing results, handling configurations) by describing what you want in plain language.
Depict will support the Model Context Protocol (MCP), so you can work with Depict through Claude, ChatGPT, Gemini, and Shopify's Sidekick, whichever assistant your team already uses. In Q4, the same natural-language configuration comes natively to the Depict dashboard, no external assistant required.
We're also shipping fully agent-readable developer documentation, so your developers' AI tools can work with our platform as fluently as your developers do.
Ecosystem and integrations
In Q4, we launch a native integration for Salesforce Commerce Cloud (SFCC), extending Depict beyond Shopify and headless setups to SFCC-based storefronts.
Published on August 5, 2026