What actions should a consultant take during setup to ensure Einstein Vision works when deployed to field users?

To ensure Einstein Vision works correctly for field users, a consultant must perform several key setup actions. First, a dataset of representative images must be uploaded and labeled (E) to serve as the baseline for training. Next, the model must be built and its quality reviewed (A) to ensure it accurately identifies objects before deployment. Finally, users must be assigned the "Einstein for Consumer Goods" permission set (C) to grant them the necessary access to use the image recognition features during their store visits.

B. Populate the custom object for Einstein optical character recognition so that the system can have a baseline to perform its AI: This is too specific to OCR and uses imprecise terminology; OCR setup involves configuration rather than populating a specific custom object with baseline data.

D. Upload a suppression data set for known competitor products so Einstein can avoid the detection: This describes a model refinement technique (negative examples), not a foundational setup action required for the feature to work initially.

Salesforce Help, Consumer Goods Cloud, Set Up Einstein Vision for Consumer Goods Cloud. This guide explicitly lists the required setup steps in order: 1. Enable Einstein Vision, 2. Assign Permission Sets to Users, 3. Create a Dataset, 4. Train Your Model. This directly supports options A, C, and E.

Salesforce Help, Consumer Goods Cloud, Create a Dataset for Einstein Vision. This document details the process of uploading images to create the necessary dataset for model training (Option E).

Salesforce Help, Consumer Goods Cloud, Train Your Model and Review Its Quality. This page explains the importance of training the model and checking the model's scorecard to validate its performance before use (Option A).