Model selection, model providers, temperature concepts, token limits, latency, and cost
Data Cloud, CRM data, knowledge, retrieval, and contextual grounding
Einstein GPT Trust Layer concepts, masking, toxicity, audit, retention, and zero-data principles
Bias, fairness, explainability, privacy, human oversight, and responsible AI review
Practical Activity
Configure or analyze a scenario involving prompt anatomy, instructions, context, grounding, constraints, output formats, and evaluation.
Build and validate an exercise covering prompt builder concepts, templates, merge fields, flows, apex, and reusable prompts and model selection, model providers, temperature concepts, token limits, latency, and cost.
Use tests, logs, reports, or documentation to verify bias, fairness, explainability, privacy, human oversight, and responsible ai review.
Key Takeaways
Explain prompt anatomy, instructions, context, grounding, constraints, output formats, and evaluation in the context of Salesforce Einstein AI Training.
Apply practical methods for prompt builder concepts, templates, merge fields, flows, apex, and reusable prompts and model selection, model providers, temperature concepts, token limits, latency, and cost.
Test, document, and troubleshoot bias, fairness, explainability, privacy, human oversight, and responsible ai review.
