Data Science Leadership Series : Part 1 – The need for AI Product Management

People keep talking about data science being such a rewarding and lucrative career, but I feel it’s time to talk about the serious gaps that’s been bugging this field, in terms of operationalizing a successful data teams. Few tech companies like the Googles and Microsofts of the world have got this working like a well-oiled machine, but majority of data science teams out there are not able get the optimal work-flow going, case-in-point the large scale layoffs see in data teams all through the tech sector. Ultimately, it’s the dollar value and revenue generated that’s attributed to the data teams, which is the most important factor, just like any software engineering team. It’s not about the cool research and blogposts that you publish through the immensely talented data scientists in your company.

Why do we see such a high rate of failure of AI projects ? I think it’s because the data teams in most companies are not given the same support teams as compared to software engineering teams. We generally see that a software product has a product team that maps user needs to business needs and takes ownership of the product, we see engineering managers who take care of the professional development needs of a software engineering team so that they can grow together with the company values and retain the talent, we see project managers who make sure the projects stay within the timeline and scope, we see quality assurance teams to support the end user product testing. But if you look at most data teams, you will see a lead data scientist or manager who themselves will be an individual contributor, managing a team of data scientists / engineers and who also manages the project.And most important of all, forget someone who actually owns the product roadmap – most data science teams don’t have a dedicated product manager, and I think that’s a recipe for failure.

Why do I say so ? Because the data scientist is expected to not just do data science and software engineering that’s related to the data product, but is also expected to manage the project, the team as well as make sure the product aligns with the business use case. But who defines the business use case ? That’s also the data scientist or someone heading that team who is also a data scientist – not someone who is a product expert. This in turn leads to business dissatisfaction, and ultimately the data team getting axed. Is it the fault of the data scientist ? Nope! It’s high time we bring some professional team structure to the data science team, ideally by borrowing from software engineering but improvising further on it.

So we come to the main topic for this article – The need for AI Product Management

An AI product manager plays a crucial role in managing and overseeing the development and deployment of AI products and solutions within an organization. Here are some reasons why having an AI product manager is important:

1. Domain Expertise: An AI product manager possesses both technical knowledge and domain expertise. They understand the intricacies of AI technology, machine learning algorithms, and data science concepts. This expertise enables them to bridge the gap between the technical teams and business stakeholders, ensuring that AI solutions align with the organization’s goals and meet customer needs.
2. Strategic Vision: An AI product manager helps define the strategic vision for AI products and solutions. They work closely with stakeholders, including executives, business leaders, and data scientists, to understand market demands, identify opportunities, and prioritize features and functionalities. Their strategic guidance ensures that AI initiatives are aligned with the overall business strategy.
3. Roadmap Planning: A dedicated product manager for data science teams is responsible for creating and maintaining the product roadmap. They collaborate with cross-functional teams, including data scientists, engineers, designers, and business analysts, to identify key milestones, set priorities, and manage dependencies. This ensures a well-structured and organized approach to AI development.
4. Requirement Gathering: Understanding user needs and translating them into technical requirements is crucial for successful AI product development. An AI product manager works closely with business stakeholders and end-users to gather requirements, elicit feedback, and validate ideas. They act as a liaison between data science teams and stakeholders, ensuring that the developed solutions address real-world problems effectively.
5. Stakeholder Management: An AI product manager facilitates effective communication and collaboration among different stakeholders. They serve as a point of contact for inquiries, feedback, and updates related to AI products. By managing expectations and fostering collaboration, they help build strong relationships between data science teams and other departments within the organization.

Disadvantages of not having a dedicated product manager for data science teams

1. Lack of Focus: Without a dedicated product manager, data science teams may face challenges in prioritizing tasks and aligning their efforts with business objectives. This can lead to a lack of clarity, confusion, and inefficient use of resources.
2. Misaligned Priorities: Data science teams may focus solely on technical aspects and lose sight of the broader business goals. Without a product manager’s guidance, there is a risk of developing solutions that do not address the right problems or deliver value to customers.
3. Ineffective Communication: A product manager acts as a communication bridge between technical teams and business stakeholders. Without their presence, there may be a breakdown in communication, leading to misunderstandings, delays, and missed opportunities.
4. Poor User Experience: Data science teams may develop technically advanced solutions but fail to address user experience aspects. A product manager ensures that user needs are considered throughout the development process, leading to more user-friendly and intuitive AI products.
5. Lack of Accountability: Without a dedicated product manager, it becomes challenging to assign clear accountability for AI product development. This can result in delays, lack of ownership, and difficulties in tracking progress.

In conclusion, if you are in data science leadership, or aspiring to be a data science manager, I feel it’s important to drill down on your AI product management situation. No matter whether it’s a small team or a large organization, investing on the AI product manager role and getting it right would solve a lot of business problems even before they pose as real risks.