Google’s AI try-on now works from a single selfie—Gemini ‘Nano Banana’ builds a full-body avatar, rolling out in the U.S.

## Morocco’s retail AI moment Morocco’s online fashion market is growing, but fit remains a hard problem. Many shoppers still rely on Instagram, WhatsApp, and in-store trust. Returns are costly, and sizing is inconsistent across brands. That is why “selfie-first” AI matters in Morocco. A headshot is easy to capture on any phone. It lowers the barrier for virtual try-on, even for casual shoppers. Morocco also has the building blocks to adopt these tools fast. The country has active tech communities in Casablanca, Rabat, and other cities. It also has public programs pushing digital transformation, plus local hubs like Technopark. ## What Google just changed Google upgraded its AI apparel try-on so it can start from a single selfie. Shoppers no longer need to upload a full-body photo to begin. The system generates full-body images that can act like a personal mannequin. Under the hood, Google uses its Gemini 2.5 Flash Image model. Google’s internal code name is “Nano Banana.” After you upload a headshot, it produces several studio-style full-body renderings. You then pick the generated image that looks most like you. You also select your usual clothing size before previewing outfits. Google says you can keep using a full-body photo if you prefer. Google is also keeping “reference models” in the flow. If you do not want to upload photos, you can choose from diverse model bodies. That matters for users who prefer privacy or do not want a personal avatar. ## Where it is available This upgrade starts rolling out in the U.S. first. For Morocco, that detail matters. Many global shopping features arrive later, and merchant support varies by country. Still, Moroccan founders and retailers should track this release now. It signals the direction of travel for shopping, search, and ad formats. It also shows how fast generative imaging is becoming a default UX. ## How to use Google’s one-selfie try-on The flow is designed to work from Google Search and Shopping surfaces. You start from a product listing that supports the feature. **Steps** - Go to **g.co/shop/tryon**. - Or open a product in **Search, Shopping, or Images** with the **“try it on”** badge. - Upload a **selfie (headshot)**. - Select your **usual size**. - Choose one generated full-body image as your **default**. - Use that default across future try-ons, powered by Google’s **Shopping Graph**. Google positions this as usable across “billions of products” in its Shopping Graph. Practically, coverage will depend on merchants, categories, and listing quality. ## Why one selfie changes the conversion equation Virtual try-on usually fails at onboarding. Full-body photos are awkward to take and to share. Many users drop off before they ever see a preview. A single selfie removes that friction. It also matches real behavior in Morocco, where selfie cameras drive most content. If onboarding is easy, more shoppers will test try-on before leaving the product page. Apparel is also a category where photos do not answer key questions. Shoppers want cues about drape, proportions, and styling. Even an imperfect preview can reduce uncertainty. ## What this signals for Morocco’s e-commerce ecosystem Even if the feature is U.S.-only today, Moroccan commerce will feel the ripple. Global platforms tend to standardize expectations. Shoppers will ask, “Why can’t I preview this on me?” This matters for Moroccan brands selling internationally. Diaspora buyers often hesitate on sizing. A “try-on” layer could improve confidence before shipping across borders. It also matters for local marketplaces and D2C sites. Many Moroccan sellers compete on visuals and trust signals. AI try-on can become a new trust cue, like reviews or delivery badges. The practical constraint is catalog readiness. Virtual try-on needs clean product images, accurate sizes, and structured data. That pushes merchants toward better product feeds and better photography. ## Practical uses Moroccan teams can build now Google’s feature is one implementation, not the only path. Moroccan startups can apply similar ideas in narrower, local-first products. The key is to target the pain points Moroccan retailers face every day. **High-impact use cases for Morocco** - **Size guidance and fit confidence** for local sizing quirks and brand variance. - **Styling previews** for outfits, bundles, and “complete the look” offers. - **Returns reduction** by setting expectations before purchase. - **Assisted selling** for WhatsApp commerce, using quick try-on previews in chats. - **Catalog normalization** tools that flag missing sizes, unclear photos, or bad descriptions. - **Creator commerce** assets for influencers, with compliant, labeled synthetic visuals. Traditional Moroccan clothing also deserves attention. Djellabas, kaftans, and modest wear have different drape and silhouette rules. A local product can focus on those garment types, where global datasets may be weaker. ## The role of government, regulation, and trust Morocco has an established data protection framework. **Law 09-08** regulates personal data processing. The **CNDP** is the authority that oversees compliance. Selfies are personal data, and often biometric-adjacent. Any Moroccan retailer or startup building “try-on” should treat this as high sensitivity. Trust will decide adoption more than model quality. **Operational best practices for Moroccan builders** - Ask for **clear consent** and explain what the image is used for. - Practice **data minimization** and keep images only as long as needed. - Provide a **delete** option that is easy to find. - Separate **analytics** from identity when possible. - Add visible labels for **AI-generated** outputs. Government digital transformation efforts can help, but procurement is not the first lever here. The near-term opportunity is capacity building and standards. Morocco needs more shared playbooks for privacy, dataset governance, and AI evaluation in Arabic and French contexts. ## Risks that Moroccan retailers should not ignore Generative try-on can mislead if it looks too “real.” Fabric behavior, transparency, and stretch are hard to simulate. Lighting and pose can hide issues that matter for fit. Bias is another concern. If generated bodies or skin tones do not represent Moroccan diversity, shoppers will notice. Offering diverse reference models and user controls helps, but it is not a full fix. There is also fraud and brand risk. AI images can be repurposed to sell counterfeit goods, or to misrepresent product quality. Retailers should watermark assets and keep source-of-truth photos visible. ## Google Doppl and the shift to immersive shopping Google is also testing a separate app called **Doppl**. It experiments with a shoppable discovery feed, with direct merchant links. It also shows AI-generated videos for product discovery. For Morocco, this is a hint about the next interface. Shopping may look more like social video, but powered by search and structured product data. Moroccan brands that already work with creators should prepare for stricter asset requirements and clearer disclosure. ## What Moroccan startups can learn from “Nano Banana” The technical headline is not just the model name. The product lesson is the workflow. Google moved complexity behind the scenes and kept inputs simple. Moroccan teams can copy that strategy. Start with one photo, one clear sizing step, and fast results. Then iterate on realism, garment coverage, and controls. The distribution lesson matters too. Google attached try-on to surfaces people already use. Moroccan startups should think the same way, via web widgets, Shopify-style plugins, or messaging integrations. ## A practical checklist for Moroccan merchants You do not need Google’s feature to start preparing. The readiness work is the same for any virtual try-on or richer product preview. **Merchant readiness checklist** - Use **consistent product photos** and avoid heavy filters. - Maintain **accurate size charts** and local size conversions. - Track **return reasons** and map them to sizing or photo issues. - Improve **structured product data** in your catalog. - Create a policy for **synthetic images** and disclose them. - Train support teams to explain what try-on **can and cannot** guarantee. If you sell to tourists or the diaspora, add localization early. Use Arabic and French sizing language. Make delivery and exchange rules explicit. ## Key takeaways - Google’s try-on now starts from a **single selfie**, using Gemini 2.5 Flash Image (“Nano Banana”). - The feature rolls out **in the U.S. first**, but it sets expectations for global e-commerce. - Morocco’s **mobile-first** shopping habits make selfie onboarding especially relevant. - Local teams can build value around **fit confidence**, catalog quality, and messaging-based assisted selling. - Privacy and consent are central, under Morocco’s **Law 09-08** and CNDP oversight. ## What to watch next The next questions are about coverage and control. Which categories and merchants will support try-on at scale. How much transparency users get about what is generated. For Morocco, the opportunity is to prepare before the format becomes table stakes. Merchants can strengthen catalogs and sizing data now. Startups can ship niche try-on products that fit local shopping behavior, then expand.