Salesforce · 2026 Edition
A complete preparation guide written by Salesforce-certified engineers. Covers the exam format,all 4 blueprint domains, a week-by-week study plan, and proven tips for passing first time.
3–5 weeks
Prep time
Beginner
Difficulty
40
Exam questions
65/1000
Pass mark
Exam code
AI Associate
Full name
Salesforce AI Associate
Vendor
Salesforce
Duration
70 minutes
Questions
40 items
Passing score
65/1000 (scaled)
Domains covered
4 blueprint domains
Recommended experience
Basic Salesforce platform familiarity; no prior AI experience required
Typical prep time
3–5 weeks
The Salesforce AI Associate validates foundational understanding of AI concepts and how Salesforce Einstein AI is used within the Salesforce platform. It is the entry-level AI credential in the Salesforce ecosystem — ideal for admins, business analysts, and consultants who need to understand, configure, and advocate for AI features in Salesforce CRM.
Job roles this opens
Domain percentage weights are not currently available for this exam. The checklist below is still useful for planning your study.
Week 1
AI Fundamentals: supervised vs unsupervised learning, classification vs regression, training data quality, model evaluation basics
Tip: This exam tests conceptual AI literacy, not technical ML skills. Know the difference between classification (predicts a category — e.g. 'Will this lead convert?') and regression (predicts a number — e.g. 'How much will this deal be worth?'). Know why training data quality matters and what bias in training data causes.
Week 2–3
Salesforce Einstein AI: Einstein Prediction Builder, Einstein Discovery, Einstein Language, Einstein Vision, Einstein Bots, Copilot
Tip: Know each Einstein product and its use case: Einstein Prediction Builder (predict custom outcomes on any Salesforce object), Einstein Discovery (analytics + recommendations), Einstein Language (NLP — sentiment analysis, intent classification), Einstein Vision (image classification), Einstein Bots (conversational AI for service). Einstein Copilot is the newest — know it uses generative AI for CRM tasks.
Week 4–5
Ethical AI & Data Governance: Salesforce Trusted AI Principles, bias mitigation, data privacy, explainability, GDPR in AI context
Tip: Salesforce's five Trusted AI Principles are tested: Responsible (minimise unintended consequences), Accountable (clear human oversight), Transparent (understandable decisions), Empowering (augments human capability), Inclusive (serves diverse users equitably). Know how each principle applies to Einstein AI features.
The Salesforce AI Associate exam (60 questions, 105 minutes, 70% pass mark) covers three main domains: AI Fundamentals (17%), Einstein AI capabilities (66%), and Ethical AI & Data (17%). Einstein capabilities is the dominant domain — focus most study time here.
Einstein Prediction Builder vs Einstein Discovery: Prediction Builder lets admins create custom predictions on any object without code (e.g., predict churn risk on Account). Discovery analyses historical data and provides statistical insights with recommendations. Know when to use each.
Data quality for Salesforce AI: Einstein models require sufficient historical data (Einstein recommends 400+ records minimum), complete required fields, and accurate labels. Incomplete or biased Salesforce data produces inaccurate Einstein predictions — this is a key exam concept.
Einstein Bots are built with the Bot Builder and handle routine service enquiries. Know the key concepts: intent (what the customer wants to do), entity (specific data extracted from the conversation, e.g. case number), dialog (conversation flow), and handoff to human agent when the bot cannot resolve the issue.
Responsible AI in Salesforce context: know that Salesforce Model Cards describe what an Einstein model does, its limitations, and its ideal use cases. Know that admins should review Einstein prediction explainability to check which fields drive predictions and whether those drivers make business sense.
Apply everything in this guide with adaptive practice questions, detailed answer explanations, and domain analytics.