Your company plans to build a generative AI solution based on internal data.
You recommend using Microsoft Foundry as a starting point to develop and manage the solution.
What is a key benefit of using Microsoft Foundry for this project?
A.Provides a scalable platform for developing and deploying generative AI solutions.Correct
B.Removes the need to select or configure the underlying AI model.
C.Enables business users to build generative AI solutions.
D.Offers a low-code platform for developing generative AI solutions.
Explanation
Microsoft Foundry is an enterprise platform that brings agents, models and tools together, with managed endpoints, scaling, identity, observability and governance for building and deploying AI solutions. It does not remove the model decision: you still choose a model from the catalog and configure it. It is aimed at developers rather than business users, who are better served by tools such as Copilot Studio. The portal can build some agents without code, but low-code development is not its defining benefit, and its SDKs and hosted code are central to the platform.
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Explanation
Statement 1 (Yes): Foundry applies Microsoft Entra identity, role-based access control, content filters, network isolation and Azure Policy across projects. Agents, models and tools are managed together under one management plane, which gives a governed environment for building and running solutions.
Statement 2 (Yes): Foundry runs agents and models as managed services, and hosted agents get a managed endpoint with scaling. Organizations can grow workloads as usage increases without building that infrastructure themselves.
Statement 3 (Yes): Foundry is not limited to text generation. Foundry Tools include Vision, which analyzes content in images and videos, and Custom Vision, which customizes image recognition for a business, and both can be tried from the Foundry portal.
Select the answer that correctly completes the sentence.
Explanation
Azure Machine Learning is the right choice for the dropdown because it trains custom predictive models from your own data, and its automated ML supports time-series forecasting for scenarios such as sales and customer demand. Azure AI Search indexes and retrieves content and does not train forecasting models. Azure OpenAI provides generative language models rather than a forecast trained on historical sales. Microsoft Foundry centers on building agents and generative AI applications, so it is not the service aimed at training this kind of tabular forecasting model.
Which business requirement most closely relates to grounding a generative AI model?
A.supporting multiple languages
B.measuring the number of user interactions per day
C.enabling users to interact by using natural language queries
D.ensuring that verified company data sources are used for response generationCorrect
Explanation
Grounding gives a model trusted data to draw its responses from, so using verified company sources is exactly that requirement. It reduces, but does not eliminate, inaccurate responses. Multiple language support and natural language interaction describe how users communicate with the model, not where its answers come from. Counting daily interactions is a usage metric and has no effect on the content of responses.
You need to create a custom Azure Machine Learning model. The data used to train the model is consistent and uniform.
What should you do first?
A.Prepare the training data.Correct
B.Evaluate the model.
C.Train the model.
D.Tune hyperparameters.
E.Deploy the model.
Explanation
A model is trained from data, so the training data must be prepared and made available first, even when it is already consistent and uniform. It still has to be selected, structured and split for training and validation. Training cannot start until that data is ready. Tuning hyperparameters and evaluating the model happen during or after training, and deployment comes last.
Your company uses a non-reasoning generative AI model to create textual content.
You discover that the model’s responses are inconsistent and do NOT meet expectations.
You need to improve the prompts.
What should you do? More than one answer choice may achieve the goal. Select the BEST answer.
A.Provide the prompts with extensive examples of the expected output.
B.Add the context, sources, and expectations to the prompts.Correct
C.Use technical terms in the prompts to enhance AI comprehension.
D.Add only a single concise requirement to the prompts.
Explanation
Inconsistent results usually mean the prompt leaves too much to interpretation, so adding context, source material and clear expectations is the best fix. Grounding data and specific instructions are among the most effective ways to get reliable output. Examples can also help, but on their own they do not supply the missing context or sources, and extensive ones are not required. Technical terms do not improve comprehension, and a single concise requirement leaves even more for the model to guess.
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Explanation
Statement 1 (No): Dataset size gives no guarantee of accuracy. Generative models can fabricate content that sounds reasonable but is wrong, and even grounding in trusted data only reduces the likelihood of inaccurate responses, so outputs still need validating.
Statement 2 (Yes): Microsoft integrates its models with guardrails, previously called content filters, and abuse detection. These safeguards help reduce harmful output, although they work best alongside monitoring and human review.
Statement 3 (No): The word always makes this false. Preparing and reviewing data helps, but models can still reinforce stereotypes and behave unevenly across groups, so outputs must be tested and monitored for fairness.
Select the answer that correctly completes the sentence.
Explanation
The dropdown is completed by generative, because creating new and original content from patterns learned in large datasets is what defines generative AI. Predictive AI also learns patterns from data, but it uses them to estimate future outcomes or values rather than produce new content. Prescriptive AI recommends actions to take based on data and likely outcomes.
Your company plans to use generative AI to help project managers and engineers work with construction blueprints stored as PDF files.
You need to recommend a generative AI solution that meets the following business requirements:
- Processes both images and text
- Summarizes the design of a building
- Answers user questions about a building's design
- Extracts information from blueprints, such as the location of electrical, heating, and plumbing systems
What should you recommend?
A.a multi-modal solutionCorrect
B.an optical character recognition (OCR) solution
C.a text completion solution
D.a document summarization solution
Explanation
A multi-modal solution accepts more than one type of input, so it can interpret the drawings in a blueprint together with its text, then summarize the design, answer questions, and locate systems such as electrical and plumbing. OCR only reads characters from an image and does not understand a drawing or answer questions about it. A text completion solution works with text alone, and a document summarization solution covers only one of the four requirements.
In which scenario is Azure Machine Learning most likely to deliver strategic value for an organization?
A.Using historical sales data to forecast demand across product categories.Correct
B.Digitizing a paper-based process to reduce errors.
C.Entering customer feedback into a spreadsheet to understand sentiment.
D.Sending personalized emails to customers based on the customer location.
Explanation
Azure Machine Learning is a service for training, deploying, and managing machine learning models, and forecasting demand from historical sales data is a typical use of a trained model. Digitizing a paper process is a data capture task that needs no predictive model. Typing feedback into a spreadsheet is manual work, and emailing customers based on location is a simple rule that does not require machine learning.
Select the answer that correctly completes the sentence.
Explanation
The primary goal of generative AI is to create new content, such as text, images, or code, usually in response to a prompt. Analyzing trends and classifying data describes analytics and classification models, which label or summarize existing data rather than produce something new. Making predictions based on historical data describes predictive machine learning such as forecasting.
Select the answer that correctly completes the sentence.
Explanation
Using high-quality grounding data in a generative AI solution improves the accuracy and reliability of the predictions and outputs of AI, because the model draws its responses from trusted data rather than only from what it learned in training. It does not ensure that all responses are factually accurate: a grounded model can still misrepresent its source, so outputs must still be validated. Grounding also does not change the speed of the model or the storage needed to host it, because the data is supplied at request time.
Select the answer that correctly completes the sentence.
Explanation
Prompt engineering is the process of crafting clear instructions to guide generative AI solutions in generating context-appropriate content, for example by stating the task, adding context and examples, and specifying the output format. Integrating tools into workflows is an adoption and integration activity. Fixing errors in generated content happens after the output exists, while prompt engineering shapes the request before it. Designing and training models is model development, and prompt engineering does not change the model itself.
Select the answer that correctly completes the sentence.
Explanation
When a generative AI model produces output that seems realistic but contains incorrect information, the behavior is known as model inaccuracy, often called fabrication or hallucination. Data leakage is the exposure of sensitive information. Overreliance describes people trusting output without checking it, so it is a user behavior rather than a model behavior. Prompt injection is an attack in which malicious instructions in the input try to make the model act in unintended ways.
Your company uses a generative AI solution.
You need to improve the quality of responses by using grounding.
Which statement accurately describes how grounding improves accuracy and relevancy?
A.references a diverse set of people, disciplines, and perspectives
B.explains how and why AI models generate content
C.anchors the responses in specific data sourcesCorrect
D.specifies the strengths and weaknesses of the AI model
Explanation
Grounding gives the model specific data to draw its responses from, such as web content or organizational data, so answers are anchored in those sources rather than only in what the model learned in training. Referencing diverse people and perspectives relates to inclusiveness, not grounding. Explaining how models generate content is transparency, and describing a model's strengths and weaknesses is model documentation; neither supplies data to the response.
You plan to meet with a group of stakeholders to discuss how generative AI can benefit your company.
You need to provide the stakeholders with a relevant description of generative AI during the meeting.
Which description should you use?
A.Generative AI is designed to generate responses based on a user's natural language prompts.Correct
B.Generative AI is designed to translate documents into other languages.
C.Generative AI is designed to predict future trends based on historical data.
D.Generative AI is designed to recommend products based on user behavior.
Explanation
Generative AI creates content and answers questions in response to prompts written in natural language, which is the description that fits its overall purpose. Translation is one task it can perform, but it does not describe the technology as a whole. Predicting trends from historical data describes predictive machine learning, and recommending products from behavior describes a recommendation system.
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Explanation
Statement 1 (No): Forecasting sales from historical data is a prediction task, which belongs to traditional, nongenerative machine learning such as a forecasting model. Generative AI creates new content in response to a request and is not designed to produce reliable numeric forecasts.
Statement 2 (Yes): Summarizing information is one of the most common uses of generative AI. A language model can condense a long policy document into its key points, although the result should still be checked against the original.
Statement 3 (Yes): Drafting new content from supplied context is what generative AI does. Given product specifications as input, a language model can write a matching product description.
Which statement accurately describes the difference between a pretrained generative AI model and a fine-tuned generative AI model?
A.A pretrained model requires labeled data, while a fine-tuned model does not.
B.A pretrained model is faster to train than a fine-tuned model because the pretrained model uses fewer parameters.
C.A pretrained model is trained on broad datasets, while a fine-tuned model is adapted to perform well on a narrower, domain-specific dataset.Correct
D.A pretrained model is optimized for a specific task, while a fine-tuned model is designed for general-purpose use.
Explanation
A pretrained model is trained on broad data so that it works for general purposes. Fine-tuning gives that pretrained model additional training on a specific task or dataset, adapting it to a domain such as medicine, finance, or law. Option A is backwards, because fine-tuning is the step that needs prepared training examples, and option D reverses the two model types. Option B is wrong because a fine-tuned model starts from a pretrained model and has the same architecture, so parameter count does not separate them.
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Explanation
Statement 1 (No): RAG and fine-tuning are separate approaches. RAG adds retrieved content to the prompt, while fine-tuning is used to change a model's behavior, style, or task performance, so RAG does not require it.
Statement 2 (Yes): A model only knows the data available when it was trained, and RAG is the recommended pattern for frequently changing information. Responses are grounded in retrieved content and can include citations back to the source, which makes them verifiable.
Statement 3 (Yes): RAG retrieves relevant passages from your own documents and supplies them to the model as grounding data. The base model is unchanged, so no retraining is needed to get answers based on your organization's content.
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Explanation
Statement 1 (Yes): People and organizations hold back from AI when they doubt that it protects their data or gives correct answers. Addressing risk and building confidence through governance and security is therefore part of adopting AI.
Statement 2 (Yes): AI output depends on the data behind it, so poor or inconsistent data leads to untrustworthy results. Clear policies for data consistency and quality are needed for AI to perform well.
Statement 3 (Yes): AI adoption is an organizational effort, not only a technology choice. Business leaders define the outcomes AI should support, and governance and compliance requirements set the boundaries, so involving both groups gives adoption a clear purpose and keeps it within policy.
Verifying access...
Access Required
You need to purchase this exam to access practice mode
Unlock a few more questions every day for 300 points you earn free on
Godlike Academy. Everything you unlock with the Daily Climb stays open for 7 days after your last unlock.