PMI Certified Professional in Managing AI (PMI-CPMAI)

A company plans to operationalize an AI solution. The project manager needs to ensure model performance is meeting selected thresholds before release.

What is an effective way to confirm these thresholds before this release?

An AI project team has completed an AI go/no-go assessment. They have discovered several technology and data factors to be insufficient. Which action should occur?

A financial services firm is assessing the success of a newly operationalized AI system for fraud detection. The project manager needs to evaluate the model against business key performance indicators. (KPIs).

What is an effective method to help ensure the accuracy of this evaluation?

Different AI project team members are responsible for various parts of the project, both cognitive and non-cognitive. The project manager needs to ensure effective accountability documentation.

Which method will help to ensure accurate documentation?

A healthcare provider had physicians review a potential diagnostic AI application. During their final review, the project team along with the physicians, discovered that the AI model exhibits a higher than acceptable false-positive rate.

Before making the go/no-go AI decision, which next step should be performed by the team?

A manufacturing company is considering implementing an AI solution to optimize its supply chain. The project manager needs to determine if AI is necessary for this task.

Which action will address the requirements?

A healthcare project manager is evaluating whether to implement an AI-powered diagnostic tool. The initial cost is US$500,000 with an expected return on investment (ROI) of 15% within the first year. The project needs to satisfy multiple stakeholders including hospital administrators and medical staff.

Which method will maximize a positive ROI for the AI implementation?

An AI project team is assessing the scalability of a healthcare solution. Which factor should the project manager consider to help ensure the solution is scalable?

A project team is currently evaluating an AI solution. They need to ensure the machine learning model provides the expected business benefits.

Which critical factor should the project manager assess?

An IT services company is working on a project to develop an AI-based customer support system. During data preparation, the project manager needs to clean and transform customer interaction logs.

What is an effective technique to handle any missing data?

A project manager is leading a complex project for a global financial institution. The project is developing an AI-driven system for real-time fraud detection and risk management. The system needs to adhere to all financial regulations. The project manager has identified skills gaps with the existing available resources.

What should the project manager do?

A company needs to launch an AI application quickly to be the first to the market. The project team has decided to use pretrained models for their current AI project iteration.

What is a key result of leveraging pretrained models?

A healthcare provider is operationalizing an AI tool to assist in diagnostic processes. To ensure robust model governance, they need to address data privacy and ethical considerations.

What should the project manager do?

A project manager is evaluating the readiness of the team for an AI operationalization project. They have identified a lack of experience with AI and data knowledge.

If this issue is neglected, what is the outcome?

An organization’s leadership team is concerned about the ethical implications of operationalizing their AI model. How should the project manager address these concerns in their presentation to the team?

A company’s leadership team has requested insights into the AI model’s ability to support decision-making processes without requiring them to understand complex technical details. Which step should the project manager take?

A financial services firm is building an AI model to detect fraudulent transactions. Identifying and validating data sources is critical to the model’s success.

What is an effective method that helps to ensure data accuracy?

A project manager is reviewing the performance of an AI model used for predictive analytics in sales. The model’s accuracy is within acceptable limits; however, its precision is low.

What is the cause for the precision issue?

A telecommunications company is preparing data for an AI tool. The project team needs to ensure the data is in the right shape and format for model training. In addition, they are working with a mix of structured and unstructured data.

Which method will address the project team’s objectives?

A project involves integrating AI systems across multiple departments, each with different access levels. This complex Al project has presented the project manager with significant issues related to data misuse. The project team has been focused on their ethics guidelines but continues to experience data misuse. The project involves different regional data protection regulations which further increases the complexity.

What issue will cause these challenges to occur?

A transportation company is preparing data for an AI model to optimize fleet management. The project team is working with large amounts of structured and unstructured data.

If the project manager avoids addressing the variety of data during preparation, what will be the result?

An organization is planning their digital transformation initiatives by building an AI solution to focus on data-collection needs. The goal is to reduce the manual handling of data.

Which approach should be prioritized to achieve the objective?

A healthcare organization is developing an Al model to predict patient outcomes. The project manager needs to address issues of missing or incorrect data before training the model.

What is an effective method to help ensure data quality?

In an aerospace manufacturing project, engineers are preparing data to train an AI system for predictive maintenance. They need to transform the data from multiple sensors and ensure it is consistent and accurate before building the model.

What should the project manager do to handle the inconsistencies?

Using machine learning and other cognitive approaches to understand how to take past / existing behavior and predict future outcomes or help humans make decisions about future outcomes using insight learned from past behavior / interactions / data is a core part to which pattern(s) of AI?

Your team is ready to operationalize the model they have been working on. It’s a model that is meant to be used on an “edge device”, specifically a mobile phone and the user may sometimes be in remote locations without regular access to the internet.

What’s the most important thing to consider here?

Your team is testing the NLP model they just created to make sure it’s performing as expected. Some of your team members want to move this model to production and move to the next iteration.

What’s wrong with this workflow?

You have been brought on to manage a recognition project, specifically an image recognition project, for an Autonomous Retail application. You know that you need to make sure you have sufficient data for this project.

What’s the best way to approach this?

Your team has collected petabytes of data for your AI project. As the project lead, you understand this is too much data to use for this iteration of the project.

What is the best course of action to take with this data?

Your company is insisting on running an automation project and applying AI best practices and methodologies to the project. You understand that automating things is just the act of using machines to repeat tasks, and does not require AI to achieve results. You think it is overkill but the project moves forward as planned.

What would likely have helped avoid this conflict?

You have been tasked at your organization to manage a large language model (LLM) project. Identify what LLMs are useful for. (Choose all that apply.)

A team is getting ready to begin working on a ML project. They need to build a data preparation pipeline and someone on the team suggests they reuse the same pipeline they created for their last project.

What’s wrong with this suggestion?

Use cognitive technologies/AI when you can’t code the rules or you can’t scale easily with people or automation. As a good rule of thumb when deciding if AI is right for the project you should:

Recently, you implemented an augmented intelligence application at work to help employees do their job better. However, employees have been resistant to this change and aren’t using the application as expected.

What could have been done better to get the team to feel comfortable with this technology and use it? (Choose all that apply.)

Your team is looking to develop an RPA bot to help with back-office processes such as data entry. What type of bot should your team be creating?

The confusion matrix measures how the algorithm performs for a binary classification activity. As your team is running tests to evaluate model performance, they are seeing the model is incorrectly categorizing flowers as trees.

Your model is provided the following:

Your team is working on a new loan decision model that takes a number of factors and data points into consideration and then automatically approves or denies a loan. After a month in operation someone does a review and notices that the system is denying a large number of loans from a certain demographic when all other factors from people in other regions (such as age, salary, and credit score) are the same.

What is most likely happening here?

You’ve built your model and now need to see if it actually works as expected. In which phase of CPMAI is this done?

During CPMAI Methodology Phase IV: Model Development, which of the following is not done during this phase?

Your team has built a new robot that roams the halls at your organization and helps with various things such as small deliveries. However, you notice that many employees are opting not to use the robot. When you ask them why they tell you that the robot looks “creepy” and they would rather not interact with it.

What’s going on here?

Your team has been asked to summarize and highlight patterns in historical purchasing data, identifying prior performance metrics and patterns. What type of analytics is most appropriate to apply for this need?

As an organization building an AI solution for your current customers based in NYC, but with possible plans for future expansion, how should you handle worldwide AI laws and regulations?

Your model is going to be used for continuous monitoring of machinery, with need for continuous, instant model predictions. What’s the most appropriate Model Operationalization approach?

Your team is working on a new facial recognition application. Since this technology has the potential to be mis-used you think it’s important to set guidelines for the proper use of this application and you want to make sure the AI system is built for some positive purpose.

What area of Trustworthy AI does this best fall under?

You are working on the data engineering pipeline for the AI project and you want to make sure to address the creation of pipelines to deal with model iteration. What part of the pipeline best deals with this step?

You’re being told by upper management that you need to manage a new AI project. You need to determine the AI project fit to make sure you’re actually solving a real business problem.

During phase I: Business Understanding, you should consider at least one of the following (Choose all that apply):

You are working on the data engineering pipeline for the AI project and you want to make sure to address the creation of pipelines to deal with model iteration. What part of the pipeline best deals with this step?

You’re testing your model and it is overly sensitive to the fluctuations of data and having trouble generalizing. What type of problem is this?

A team has started working on their first AI project and they are running this project like a traditional software development project. About two months into the project the team is hitting some major issues, and you’re tasked with coming in to help manage this project. Immediately you realize that AI projects need to be treated like data centric projects.

What’s the next best course of action?

You are establishing the data requirements for the project. Which of the following tasks is the least likely to impact data requirements?

You have been tasked with creating a model that will recommend products based on what other customers have similarly purchased. Which algorithm is the best choice given this situation?

You’re working with an inexperienced team and this is all their first AI project. You’re trying to work on a supervised learning binary classification problem to determine if emails are spam or not.

What is the best approach for this project?

In order for Supervised Learning approaches to work, they must be fed clean, well-labeled data that the system can use to learn from examples. But how do you get Labeled Data?

As a team leader at a small startup, what approach would not be beneficial when trying to gather labeled data?

Creating machine learning models can be complicated. Your team wants to use tools called Automated Machine Learning (AutoML) to simplify the process. You know of another team that has used AutoML tools and it’s saved the team a lot of time.

However, what’s the one area you should not have the AutoML tool help with?

One of the key elements of a data-centric methodology is the data requirements phase. During CPMAI Phase II, several unexpected issues have developed and are now threatening the data collection efforts.

What course of action might make the issue worse?

The team is working to build a data preparation pipeline for the conversational chatbot project. Which phase of CPMAI is this done?

For AI projects the code and systems don’t matter as much as the data. In fact, big data is what’s powering much of this latest wave of AI. What’s most important for your company to consider around data?

The growth of Big Data has led to a desire to be able to do more to process and extract more value from Big Data. Simply storing data and providing analytics is no longer enough anymore to remain competitive.

To keep your organization competitive, you need to: