To provide your customers with more ways to find products, Visualsoft has developed Image Search using the Google Cloud Vision API. Image Search allows your customers to take photos of a product they would like to purchase from you e.g. a celebrity in a news article or competitor's product, upload it to your site through our image search panel in the search bar, and find products that you sell that are similar. Of course, they will see the results in the search results page layout that they are familiar with already to ensure the customer’s journey is not impacted.
The UX has been designed by our in-house UX Optimisation team by looking at what your customers will have already experienced in this area with providers like Google. By looking at the journey a customer may expect, we have been able to build an image search user experience which will feel natural and familiar.
Which sectors can use this?
Currently, we fully support:
Homegoods such as furniture, appliances, and furnishings.
Fashion such as clothing, accessories and shoes.
Are you not in one of these sectors but are interested? Please contact your Success Manager and we can analyse whether we believe this functionality will work for you.
How it works
If you chose to use this feature, we use your product images to train an AI Computer Vision model which can recognise unique features of each of your products.
With the model training complete, users can use the image search button available in the search bar of your store.
Product Image Requirements
Our AI requires that product images are on a white background. This doesn’t mean all product images need to be on a white background, but 2-3 for each product should be. The more angles of your product you have, the more accurate the image search will be.
Please note that to enable this feature you need an active Google Cloud account and will be responsible for all associated costs for image searches (payable directly to Google based on usage). For further information on this, please contact your eCommerce Strategist.