Build AI-Powered Apps Fast (Pilot)

With the Predictive Vision Service, developers can harness the power of image recognition to build AI-powered apps. Without a data science degree! You can train deep-learning models to recognize and classify images at scale with the Predictive Vision Service REST API. Leverage pre-trained classifiers, or train your own custom classifiers to solve a vast array of specialized image-recognition use cases. Developers can bring the power of image recognition to CRM and third-party applications so that end users across sales, service, and marketing can discover new insights about their customers and predict outcomes that lead to smarter decisions.
Available in: All Editions except



We provide the Predictive Vision Service to selected customers through a pilot program. The Predictive Vision Service isn’t generally available unless or until Salesforce announces its general availability in documentation or in press releases or public statements. We can’t guarantee general availability within any particular time frame or at all. Make your purchase decisions only on the basis of generally available products and features. You can provide feedback and suggestions for the Predictive Vision Service on the IdeaExchange in the Success Community.

For information about signing up for the service and using the API, see the Predictive Vision Service (Pilot) Developer Guide.

When to Use the Predictive Vision Service

Images contain contextual clues about all aspects of your business, including your customers’ preferences, your inventory levels, and the quality of your products. You can use these clues to enrich what you know about your sales, service, and marketing efforts to gain new insights about your customers and take action. The possibilities are limitless with applications that include:
  • Visual search—Expand the ways that your customers can discover your products and increase sales.
    • Provide customers with visual filters to find products that best match their preferences while browsing online.
    • Allow customers to take photos of your products to discover where they can make purchases online or in-store.
  • Brand detection—Monitor your brand across all your channels to increase your marketing reach and preserve brand integrity.
    • Better understand customer preferences and lifestyle through their social media images.
    • Monitor user-generated images through communities and review boards to improve products and quality of service.
    • Evaluate banner advertisement exposure during broadcast events to drive higher ROI.
  • Product identification—Increase the ways that you can identify your products to streamline sales processes and customer service.
    • Identify product issues before sending out a field technician to increase case resolution time.
    • Discover which products are out of stock or misplaced to streamline inventory restocking.
    • Measure retail shelf-share to optimize product mix and represent top-selling products among competitors.

Classifying Images

Use the API to classify images by:
  • Accessing them via a URL
  • Uploading them from a local location
  • Passing in the image converted into a Base64 string
You can also classify images stored in Salesforce by converting them into a Base64 string.
The Predictive Vision Service is a REST API, so you can call it from any language that supports REST calls. This cURL call passes an image of a tree frog into the pre-built model called GeneralImageClassifier.
curl-X POST -H "Authorization: Bearer <TOKEN>" -H "Cache-Control: no-cache" -H "Content-Type: multipart/form-data" -F "sampleLocation=" -F "modelId=GeneralImageClassifier"
Here’s what the prediction results look like.
  "probabilities": [
      "label": "tree frog, tree-frog",
      "probability": 0.7963114976882935
      "label": "tailed frog, bell toad, ribbed toad, tailed toad, Ascaphus trui",
      "probability": 0.1978749930858612
      "label": "banded gecko",
      "probability": 0.001511271228082478
      "label": "African chameleon, Chamaeleo chamaeleon",
      "probability": 0.0013212867779657245
      "label": "bullfrog, Rana catesbeiana",
      "probability": 0.0011536618694663048
  "object": "predictresponse"