Project management for AI | TechRepublic

Project management for AI | TechRepublic


AI project management requires a different approach than traditional IT project management. What are these differences and how can an AI project be managed to succeed?

Developers working on a project
Image: nd3000, Getty Images / iStockphoto

In 2019, the number of AI projects that ultimately failed was approximately 85%, and 96% of organizations reported having problems with data quality, data labeling, and creation. of trust in the model. It was also reported that senior management was unaware of the artificial intelligence and the value it could offer.

Today, AI (and AI projects) are still in the early stages of deployment. If companies use AI, they use it in prefabricated systems from external vendors where vendors have developed AI, not their customer-companies.

In the future, however, more companies will find a reason to develop their own internal AI, and this means defining a project management approach that works with AI.

How does an AI project differ from traditional projects?

In traditional project management, even if done with methodologies such as Agile, the success of the project is defined by the software that is produced and a well-understood process. Although the development of the project is not done in a linear way like Agile, the basic steps are still to define, design, develop, test and deploy. The data with which these applications operate is almost always a structured system of log data that is already evaluated for quality and quite mature in its form and substance.

Because the data with which traditional software development operates is reliable and because everyone understands the development steps used in the project, there is much less uncertainty in traditional software development projects. This makes it possible to attach credible project deadlines based on the history of the past project.

Unfortunately, AI projects do not have this same stability, nor is it so easy to assign hard deadlines for project completion.

I’LL SEE: Contracting kit: Project manager (TechRepublic Premium)

Navigating through uncertainty in AI projects

There is no absolute “end” for an AI project, unless it’s a project you’re offline.

If you are an AI project manager, you have to live with this “endless” reality, and so do your project management and sponsors.

Why isn’t there an end?

Because AI asks questions about the data it analyzes based on the data it operates on, and that data is constantly changing. As you add new data sources, the results will change. The AI ​​itself will also include machine learning (ML) that recognizes data patterns and learns from those patterns. This can also change the results.

Your management and your users should have an understanding (and an expectation) that as the data changes, so can the results. Part of this process includes accepting uncertainty as part of the evolution of the AI ​​system.

Defining the delivery of your AI project

At some point from a project perspective, an AI project should be considered finished.

The goal of most AI projects is to achieve at least 95% compliance with AI results with what the experts in the field conclude. Once this 95% threshold is reached, the project is considered accurate enough to be launched. This is when the project must be completed.

This does not mean that all work on the resulting application or AI systems is complete. There will be
“drifts” over time that could cause AI to lose some of its accuracy. At this point, it will be necessary to recalibrate the AI ​​to provide once again with optimal quality, but this is software maintenance.

I’LL SEE: Top Keyboard Shortcuts You Need to Know (Free PDF) (Technological Republic)

Do AI project deliveries always come out as planned?

The answer is a resounding “No!”

There are times when the data used by AI is not properly prepared, especially when new and unknown data sources are introduced. Gross data will distort AI results.

Second, if your business case changes (and the value that users want to get out of it), AI will no longer adapt to what the company wants. Finally, there are only cases where AI projects don’t work, no matter how much you do. to prove. This possibility should be discussed in advance with management, and everyone should be on board to “disconnect” as soon as an AI project proves unsuccessful.



Source link

Related post

EDUCAUSE 2022: How Data Collection Can Improve Student and Faculty IT Support

EDUCAUSE 2022: How Data Collection Can Improve Student and…

At Indiana University, Gladdin said, to make life easier for students and faculty, they implemented a course template for the Canvas…
UGC, AICTE warn students against online PhD programmes offered by EdTech platforms | Latest News India

UGC, AICTE warn students against online PhD programmes offered…

The University Grants Commission (UGC) and the All India Council for Technical Education (AICTE) on Friday issued a joint advisory against…
UGC, AICTE warn students against online PhD programmes offered by EdTech platforms | Latest News India

UGC, AICTE warn students against online PhD programmes offered…

The University Grants Commission (UGC) and the All India Council for Technical Education (AICTE) on Friday issued a joint advisory against…

Leave a Reply

Your email address will not be published.