How to Pass Salesforce Certified AI Associate Exam.
The Salesforce Certified AI Associate should be able to provide informed strategies and guide stakeholder decisions based on Salesforce’s Trusted AI Principles. Candidates should be familiar with data management, security considerations, common business and productivity tools, and Salesforce Customer 360.
1. About the Salesforce Certified AI Associate Exam
|Time allotted||70 minutes|
|Passing score||65% (26 out of 40 questions)|
|Exam Fee||USD 75 plus applicable taxes|
For up to date information about this certification please refer to the official exam guide here!
2. Salesforce Salesforce Certified AI Associate Exam Outline
2.1 AI Fundamentals: 17%
- Explain the basic principles and applications of AI within Salesforce.
- Differentiate between the types of AI and their capabilities.
2.2 AI Capabilities in CRM: 8%
- Identify CRM AI capabilities.
- Describe the benefits of AI as they apply to CRM.
2.3 Ethical Considerations of AI: 39%
- Describe the ethical challenges of AI (e.g., human bias in machine learning, lack of transparency, etc.).
- Apply Salesforce’s Trusted AI Principles to given scenarios.
2.4 Data for AI: 36%
- Describe the importance of data quality.
- Describe the elements/components of data quality.
3. Salesforce Salesforce Certified AI Associate Exam Study Course
- Trailmix: Prepare for Your Salesforce AI Associate Credential
- Cert Prep: Salesforce AI Associate
- Free Practice Questions
4. Important Topics for the Salesforce Certified AI Associate Exam
2.1 AI Fundamentals: 17% ( 7 Questions)
- Main Types of AI Capabilities
- Numeric Predictions
- Robotic Navigation
- Language Processing
- Neural networks are tools for training AI models.
- Training AI by adding extra layers to find hidden meaning in data is called deep learning.
- Most important components of AI:
- Natural language understanding (NLU) refers to systems that handle communication between people and machines.
- Natural language processing (NLP) is distinct from NLU and describes a machine’s ability to understand what humans mean when they speak as they naturally would to another human.
- Named entity recognition (NER) labels sequences of words and picks out the important things like names, dates, and times. NER involves breaking apart a sentence into segments that a computer can understand and respond to quickly.
- Deep learning refers to artificial neural networks being developed between data points in large databases.
- Generative AI is technology that takes a set of data and uses it to create something new – like poetry, a physics explainer, an email to a client, an image, or new music – when prompted by a human.
- Generative AI uses two types of deep learning models:
- GANs are made up of two neural networks: a generator and a discriminator. The two networks compete with each other, with the generator creating an output based on some input and the discriminator trying to determine if the output is real or fake. The generator then fine-tunes its output based on the discriminator’s feedback, and the cycle continues until it stumps the discriminator.
- Transformer models, like ChatGPT, (which stands for Chat Generative Pretrained Transformer), create outputs based on sequential data (like sentences or paragraphs) rather than individual data points. This approach helps the model efficiently process context and is why it’s used to generate or translate text.
- Common Concerns About Generative AI:
- Hallucinations: Predictions from generative AI that diverge from an expected response, grounded in facts, are known as hallucinations. They happen for a few reasons, like if the training data was incomplete or biased, or if the model was not designed well.
- Data security: Businesses can share proprietary data in the generative AI lifecycle. Companies that offer AI services must demonstrate that trust is paramount and that data will always be protected.
- Plagiarism: LLMs and AI models for image generation are typically trained on publicly available data. There’s the possibility that the model will learn a style and replicate that style.
- User spoofing: Fake users generated with AI can interact with real users in a very realistic way. That makes it hard for businesses to identify bot networks that promote their own bot content.
- Sustainability: The computing power required to train AI models is immense, and the processors doing the math require a lot of actual power to run. As models get bigger, so do their carbon footprints.
- Generative AI Glossary
- Predictive AI studies historical data, identifies patterns and makes predictions about the future that can better inform business decisions.
|Generative AI||Predictive AI|
|Generative AI combines algorithms and deep learning neural network techniques to generate content that is based on the patterns it observes in other content.||Predictive AI studies historical data, identifies patterns and makes predictions about the future that can better inform business decisions.|
|Used for producing fresh content such as Text, Images, Product Design, Personalization||Used for financial forecasting, fraud detection, healthcare and marketing.|
- Natural language processing (NLP), is a field of artificial intelligence (AI) that combines computer science and linguistics to give computers the ability to understand, interpret, and generate human language in a way that’s meaningful and useful to humans.
- Generative AI Terms
|Generative AI Terms||What it means for customers||What it means for teams|
|Artificial intelligence (AI)||AI can help your customers by predicting what they’re likely to want next, based on what they’ve done in the past.||AI helps your teams work smarter and faster by automating routine tasks.|
|Artificial neural network (ANN)||Customers benefit in all sorts of ways when ANNs are solving problems and making accurate predictions.||Teams can forecast customer churn, which prompts proactive ways to improve customer retention.|
|Augmented intelligence||Augmented intelligence lets a computer crunch the numbers, but then humans can decide what actions to take based on that information.||Augmented intelligence can help you make better and more strategic decisions.|
|CRM with generative AI||A CRM gives customers a consistent experience across all channels of engagement, from marketing to sales to customer service and more.||A CRM helps companies stay connected to customers, streamline processes, and improve profitability.|
|Deep learning||Deep learning-powered CRMs create opportunities for proactive engagement. They can enhance security, make customer service more efficient, and personalize experiences.||In a CRM system, deep learning can be used to predict customer behavior, understand customer feedback, and personalize product recommendations.|
|Discriminator (in a GAN)||Discriminators in GANs are an important part of fraud detection, so their use leads to a more secure customer experience.||Discriminators in GANs help your team evaluate the quality of synthetic data or content. They aid in fraud detection and personalized marketing.|
|Ethical AI maturity model||Helps build trust and assures your customers that you are using their data in responsible ways.||Regularly evaluating your AI practices and staying transparent about how you use AI can help you stay aligned to your company’s ethical considerations and societal values.|
|Explainable AI (XAI)||If an AI system can explain its decisions in a way that customers understand, it increases reliability and credibility.||XAI can help employees understand why a model made a certain prediction. Not only does this increase their trust in the system, it also supports better decision-making and can help refine the system.|
|Generative AI||Better and more targeted marketing content, which helps them get exactly the information they need and no more.||Faster builds for marketing campaigns and sales motions, plus the ability to test out multiple strategies across synthetic data sets and optimize them before anything goes live.|
|Generative adversarial network (GAN)||They allow for highly customized marketing that uses personalized images or text.||They can help your development team generate synthetic data when there is a lack of customer data.|
|Generative pre-trained transformer (GPT)||Customers have more personalized interactions with your company that focus on their specific needs.||GPT could be used to automate the creation of customer-facing content, or to analyze customer feedback and extract insights.|
|Generator||Using generators, it’s possible to train AI chatbots that learn from real customer interactions, and continuously create better and more helpful content.||Generators can be used to create realistic datasets for testing or training purposes. This can help your team find any bugs in a system before it goes live, and let new hires get up to speed in your system without impacting real data.|
|Hallucination||When companies monitor for and address this issue in their software, the customer experience is better and more reliable.||Quality assurance will still be an important part of an AI team. Monitoring for and addressing hallucinations helps ensure the accuracy and reliability of AI systems.|
|Large language model (LLM)||Personalized chatbots that offer human-sounding interactions, allowing customers quick and easy solutions to common problems in ways that still feel authentic.||Teams can automate the creation of customer-facing content, analyze customer feedback, and answer customer inquiries.|
|Machine learning||When a company better understands what customers value and want, it leads to enhancements in current products or services, or even the development of new ones that better meet customer needs.||Machine learning can be used to predict customer behavior, personalize marketing content, or automate routine tasks.|
|Machine learning bias||Working with companies that actively engage in overcoming bias leads to more equitable experiences, and builds trust.||It’s important to check for and address bias to ensure that all customers are treated fairly and accurately. Understanding machine learning bias and knowing your organization’s controls for it helps your team have confidence in your processes.|
|Model||The model can help customers get much more accurate product recommendations.||This can help teams to predict customer behavior, and segment customers into groups.|
|Natural language processing (NLP)||NLP allows customers to interact with systems using normal human language rather than complex commands.||NLP can be used to analyze customer feedback, power chatbots, or automate the creation of customer-facing content.|
|Prompt engineering||When your generative AI tool gets a strong prompt, it’s able to deliver a strong output. The stronger, more relevant the prompt, the better the end user experience.||Can be used to ask a large language model to generate a personalized email to a customer, or to analyze customer feedback and extract key insights.|
|Sentiment analysis||Customers can offer feedback through new channels, leading to more informed decisions from the companies they interact with.||Sentiment analysis can be used to understand how customers feel about a product or brand, based on their feedback or social media posts, which can inform many aspects of brand or product reputation and management.|
|Supervised learning||Increased efficiency and systems that learn to understand their needs via past interactions.||Can be used to predict customer behavior or segment customers into groups, based on past data.|
|Transformer||Businesses can enhance the customer service experience with personalized AI chatbots.||Transformers help your team generate customer-facing content, and power chatbots that can handle basic customer interactions.|
|Unsupervised learning||When we uncover hidden patterns or segments in customer data, it enables us to deliver completely personalized experiences.||Teams get valuable insights and a new understanding of complex data. It enables teams to discover new patterns, trends, or anomalies that may have been overlooked, leading to better decision-making and strategic planning.|
|Validation||Better-trained models create more usable programs, improving the overall user experience.||Can be used to ensure that a model predicting customer behavior or segmenting customers will work as intended.|
|Zone of proximal development (ZPD)||When your generative AI is trained properly, it’s more likely to produce accurate results.||Can be applied to employee training so an employee could learn to perform more complex tasks or make better use of the CRM’s features.|
2.2 AI Capabilities in CRM: 8% ( 3 Questions)
- Einstein allows all Salesforce users to:
- Discover insights that bring new clarity about your company’s customers.
- Predict outcomes so your users can make decisions with confidence.
- Recommend the best actions to make the most out of every engagement.
- Automate routine tasks so your users can focus on customer success.
- Generate tailored content – from emails, to knowledge articles, to code.
- Einstein Discovery enables you to:
- Identify, surface, and visualize insights into your business data.
- Predict future outcomes and suggest ways to improve predicted outcomes in your workflows.
- Einstein Discovery-powered solutions address these use cases:
- Regressions for numeric outcomes represented as quantitative data (measures), such as currency, counts, or any other quantity.
- Binary classification for text outcomes with only two possible results. These are typically yes or no questions that are expressed in business terms, such as churned or not churned, opportunity won or lost, employee retained or not retained, and so on.
- Multiclass classification for text outcomes with 3 to 10 possible results. For example, a manufacturer can predict, based on customer attributes, which of five service contracts a customer is most likely to choose.
- You can use Einstein Prediction Service to:
- Get predictions on your data.
- Get suggested actions to take to improve predicted outcomes.
- Manage prediction definitions and models that are deployed in Salesforce.
- Manage bulk scoring jobs.
- Manage model refresh jobs.
- A prediction is a derived value, produced by a model, that represents a possible future outcome based on a statistical understanding of past outcomes plus provided input values (predictors).
- An outcome is the business result you are trying to understand and improve. An outcome is typically a key performance indicator (KPI), such as sales margin or opportunity wins.
- A prediction represents an output value that the model generates based on the provided input values (predictors). The model’s equation is the result of a thorough statistical analysis of past data with known outcomes, powered by machine learning and AI.
- A model is the sophisticated, custom mathematical construct that Einstein Discovery generates. Einstein Discovery uses a model to predict an outcome.
- A model organizes data by variables. A variable is a category of data. It’s analogous to a column in a CRM Analytics dataset or a field in a Salesforce object. A model has inputs (predictor variables) and outputs predictions for the outcome variable, along with additional information if requested.
- Predictions occur at the observation level. An observation is a structured set of data. It’s analogous to a populated row in a CRM Analytics dataset or a record in a Salesforce object.
2.3 Ethical Considerations of AI: 39% ( 16 Questions)
- Salesforce’s Trusted AI Principles
- Responsible – Safeguard human rights and protect the data we are entrusted with.
- Accountable – Seek and leverage feedback for continuous improvement.
- Transparent – Develop a transparent user experience to guide users through machine driven recommendations. Be transparent about how AI is built.
- Empowering – Promote economic growth and employment for our customers, their employees and society as a whole. Empower customers to use AI responsibly.
- Inclusive – Respect the societal values of all those impacted, not just those of the creators.
- Five guidelines Salesforce is using to guide the development of trusted generative AI
- Accuracy – Deliver verifiable results that balance accuracy, precision, and recall in the models by enabling customers to train models on their own data. Communicate when there is uncertainty about the veracity of the AI’s response and enable users to validate these responses.
- Safety – Make every effort to mitigate bias, toxicity, and harmful output by conducting bias, explainability, and robustness assessments, and red teaming.
- Honesty – When collecting data to train and evaluate our models, we need to respect data provenance and ensure that we have consent to use data. Be transparent that an AI has created content when it is autonomously delivered.
- Empowerment – Identify the appropriate balance to “supercharge” human capabilities and make these solutions accessible to all.
- Sustainability – Develop right-sized models where possible to reduce our carbon footprint.
- Ethical AI Practice Maturity Model
- Ad Hoc
- Organized & Repeatable
- Managed & Sustainable
- Optimized & Innovative
- Biggest perceived risks of real-time personalization in marketing:
- Security events, like data breaches
- Data being collected, shared, or used in unanticipated ways
- Personalizing interactions that feel invasive or unwanted to consumers
- Inadvertent bias introduced by relying on demographic attributes for interactions instead of behavioral and engagement data
- Responsible Marketing Principles when implementing personalization
- Be Transparent About Security – Being transparent about partnership with your customers can help you cement their trust in–and loyalty to—your brand.
- Develop Trust Through Consent and Transparency – When building personalized experiences, make sure that you check and obtain consent before collecting data.
- Use Data to Personalize the Experience – Providing better, more relevant experiences and, most importantly, ones that are driven by your customers’ intent. Within a curated experience, customers are more likely to share personal and sometimes sensitive information when the benefit to them is clear and understood.
- Prioritize Behavior-Based Intent Over Demographic Attributes – demographic targeting causes bias and fails to deliver the right messaging to the right people.
2.4 Data for AI: 36% ( 14 Questions)
- Factors that determine data quality
- Missing Records
- Duplicate Records
- No Data Standards
- Incomplete Records
- Stale Data
- Bad data is consistently linked with:
- Lost revenue
- Missing or inaccurate insights
- Wasted time and resources
- Slow info retrieval
- Poor customer service
- Reputational damage
- Decreased adoption by reps
- Good data lets your company:
- Prospect and target new customers
- Identify cross-sell and upsell opportunities
- Gain account insights
- Increase efficiency
- Retrieve the right info fast
- Build trust with customers
- Increase adoption by reps
- Plan and align territories better
- Score and route leads faster
- Data Quality Dimensions
- Age – What was the last time each record was updated?
- Completeness – Are all key business fields on records filled in?
- Accuracy – Has it been matched against a trusted source?
- Consistency – Is the same formatting, spelling, and language used across records?
- Duplication – Are records and data duplicated in your org?
- Usage – Is your data being harnessed in reports, dashboards, and apps?
- Standards for creating, processing, and maintaining data
- Naming – Set naming conventions for records.
- Formatting – Figure out how dates and money are represented.
- Workflow – Determine processes for record creation, reviewing, updating, and archiving. Determine all the stages a record goes through during its life cycle.
- Quality – Set appropriate standards for data quality, including the ability to measure or score records.
- Roles and Ownership – Determine the appropriate levels of privacy for data. Make sure to comply with regulatory, legal, and contractual obligations.
- Monitoring – Outline a process for ensuring quality control of data.
- Steps to clean your data to make it AI ready:
- Remove duplicate or irrelevant observations – Duplication happens when you combine data sets from multiple places, and duplicate entries are created.
- Fix structural errors – This happens when data includes typos, incorrect capitalization, or mislabelings.
- Filter unwanted outliers – There are often one-off observations that don’t appear to align with the data you’re analyzing. That might be the result of incorrect data entry.
- Handle missing data – Eliminate observations that include missing values, input missing values based on other observations or consider altering the way the data is used to effectively navigate the missing values.
- Validate – Does the data make sense, does the data follow the appropriate rules for its field?
- Data Management Best Practices
- Build a Data Management Strategy – A clear data strategy helps keep your organization on track. It’s a great way to align your team on how data will be collected, reviewed, and used to drive your business goals.
- Improve Data Quality – Tracking, reporting, and the effectiveness of your Salesforce deployment depend on getting clean data.
- Import Data – Bring your existing data into Salesforce so you can include past records in your tracking and reporting.
- Maintain and Clean Up Data – Remove duplicates.