Ronald Schmelzer and Kathleen Walch from Cognilytica discuss the importance of responsible project management in the success of AI projects. They talk about the origins of Cognilytica and the need for AI education. They also explore patterns in AI, the challenges of delivering AI projects, and the importance of aligning expectations. Collaboration and learning are highlighted as crucial factors in AI projects.
Understanding why AI projects fail is crucial for project managers, with common reasons including issues with data quality, overpromising on what AI can do, and failure to follow best practices.
Successful AI project management requires a data-centric perspective, focusing on utilizing data-specific methodologies and practices, addressing data-related challenges early on, and leveraging best practices.
Leadership buy-in, setting realistic expectations, and engaging with end-users are essential for overcoming resistance and misaligned expectations in AI projects, building trust, and minimizing resistance.
Deep dives
Understanding the Impact of AI Projects
Artificial intelligence (AI) projects have become a hot topic, but they have been around for decades. AI is all around us, from suggesting email phrasing to recommending videos and optimizing routes. However, delivering successful AI projects is a big responsibility. Not every project succeeds, and understanding why AI projects fail is crucial for project managers. Common reasons include issues with data quality, overpromising on what AI can do, and failure to follow best practices. Setting realistic expectations, scoping projects correctly, and focusing on fundamental education can increase the chances of AI success.
The Importance of a Data-Centric Perspective
AI projects are distinct in that they heavily rely on data. Unlike other digital projects, AI systems are entirely dependent on data for their functioning. Understanding the importance of a data-centric perspective is crucial for successful AI project management. This perspective means having a data mindset and utilizing data-specific methodologies and practices. Many AI projects fail due to a lack of focus on data. It is essential to address data-related challenges early on, such as accessing and preparing quality data. Following a step-by-step approach and leveraging best practices significantly improve the chances of AI project success.
Addressing Resistance and Aligning Expectations
Resistance and misaligned expectations can hinder the progress of AI projects. To overcome these challenges, leadership buy-in is crucial to address fears and concerns related to AI. Setting realistic expectations about what AI can and cannot do is essential. Organizations should focus on augmenting human capabilities, rather than replacing jobs entirely. Engaging with end-users and gathering feedback to identify pain points and areas for improvement is vital. By listening to the concerns of the team and aligning AI initiatives with their needs, trust can be built, and resistance can be minimized.
Importance of Data Access and Cleaning in AI Projects
The success of AI projects hinges upon the ability to access and clean the data. Without quality data, projects can be delayed or fail altogether. For example, if medical image data is protected by privacy regulations or is of low quality, it can hinder the project's progress. In such cases, additional efforts are required to clean the data, collect new data, or enhance existing data. This can result in unexpected time and resource requirements, throwing the project off course. Therefore, it is crucial to prioritize data access and cleaning from the start to ensure project success.
The Agile and Iterative Approach to AI Projects
The podcast emphasizes the need for an agile and iterative approach to AI projects. By starting small and iterating often, risks can be mitigated and projects can stay on track. This approach allows for flexibility and adaptation as market conditions or project needs change. Instead of aiming for a fully autonomous solution right off the bat, it is suggested to begin with smaller, focused iterations using smaller data sets or simpler models. This not only helps in managing time and costs effectively but also provides valuable insights and opportunities for improvement throughout the project lifecycle. By intertwining the different stages and components of the project, such as data understanding, data prep, model development, and evaluation, an agile methodology can be effectively applied to AI projects for increased success.
Galen Low is joined by Ronald Schmelzer and Kathleen Walch—the managing partners and lead analysts behind Cognilytica, and the hosts of AI Today—to lift the lid on why AI projects fail and why responsible project management is so important to the success of artificial intelligence.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
