Skill Sets Needed to Thrive as an AI/ML Product Manager

By Nitin Baliga on

December 19, 2024

Thrive as an AI ML Product Manager Customer focus. The technology won’t matter if the products fail to impact customers meaningfully. Data scientists often find themselves laser-focused on the latest and greatest technology, and PMs can fall into this same trap, losing sight of the ultimate goal—solving a customer’s problem. The technology and eventual solution may be the show’s star, but an effective PM will step away from those bright lights to ask the tough question: “Is this the right solution for the customer?”

Data-driven, metrics thinking. Understanding key success metrics is vital. Successful PMs approach problem-solving with a data-driven mindset. This includes breaking down the problem into areas of financial (revenue and margin), customer (active users and engagement), product (time to market, number of features developed), process (lifecycle checklists, levels of consistency, and standardization), and adoption (percentage of objectives and key results [OKR] completion, level of best practice adoption) metrics. It’s crucial for PMs to look at problems and opportunities through a broader lens with the aid of representative data while still utilizing anecdotal examples to help triangulate.

Technical skills. PMs need a deeper comprehension of technology’s application in ML and AI areas such as search and recommendations. Teams build stronger products when PMs and data engineers mutually understand and appreciate each other, so a working understanding of ML is essential.

Communication skills. Effective communication involves simplifying complex technical concepts. In that way, they can be easily understood by all stakeholders while showing empathy and appreciation for the issue’s complexity. If people don’t understand the concept, they won’t appreciate it. PMs who excel in this area facilitate more rapid innovation and impactful products.

These skills can be developed through various methods. For instance, simulating customers’ experiences (known as dogfooding) is an effective method to build empathy. For metrics, dig into the data to find patterns across segments of user data in terms of quality and engagement. Develop technical skills by taking courses that help make understanding neural networks and other ML concepts easier. While PMs aren’t typically expected to code, understanding model inputs and pressure-testing assumptions is crucial for effective problem-solving.

Communication often distinguishes a good PM from an exceptional one. Success is often defined by a person’s ability to take a complex ML concept and make it accessible to a broader audience.

Real-World Examples and Preparing for the Future

Successful product management is exemplified by industry leaders who focus on solving specific organizational challenges. Google was one of the first companies to define product management roles clearly, while Amazon famously stated its intention to be “Earth’s Most Customer-Centric Company”—one of the key tenets of product management. Other organizations emphasize the ability to balance different customer profiles while continuing to exceed expectations, such as Uber (managing drivers and consumers) and Doordash (managing drivers, customers, and restaurants).

These companies identified and met consumer needs in ways that transformed their industries. They created a market in areas that didn’t exist or expanded an existing industry beyond customers’ expectations or imaginations.

As AI and ML continue to evolve, they drive even greater opportunities for identifying customer solutions. Over the next decade, leveraging these tools to simulate experiences and build prototypes will become increasingly important. AI and ML will also improve PM efficiency by streamlining tasks such as writing compelling product requirements documents (PRDs) or summarizing notes and presentations. Additionally, they will increase opportunities for cross-departmental collaboration, strengthening communication between teams.

Navigating Challenges

In addition to re-skilling, organizations face the challenge of transitioning from predictable and definitive constructs to a more probabilistic setup. While this approach may offer greater accuracy, they are inherently more difficult to debug due to their uncertainty. The effectiveness of ML models depends heavily on the data used to train them. As such, the investment and availability of good-quality data is critical.

Developing AI- and ML-based products requires investing in data scientists and ML engineers with different talent profiles and specialized skills. Attracting this talent requires organizations to create compelling value propositions regarding compensation, culture, and team dynamics.

As the importance and implications of AI grows, the emphasis will shift from creating ideal solutions to effectively aggregating and utilizing data from those models. While technical expertise is valuable, PMs who excel in other areas can thrive by asking the right questions, understanding the implications of different solutions, communicating results effectively, and prioritizing products and features wisely.

About the Author:

Nitin Baliga is a senior director of product with more than 15 years of experience in product management, retail, strategy consulting and technology. He currently leads a team of product managers and is responsible for the quality of search results in e-commerce. Outside of search, he coordinated holiday deals and seasonal events experiences for customers and has also held leadership positions at several top-tier companies, including McKinsey & Company and Oracle. He holds an MBA degree from University of Michigan’s Ross School of Business. Connect with Nitin on LinkedIn

Disclaimer: The author is completely responsible for the content of this article. The opinions expressed are their own and do not represent IEEE's position nor that of the Computer Society nor its Leadership.