How Can We Define a Data & AI Product?
Or the Art of Adopting a Product Mindset for Data & AI Initiatives
This article is the second in a series on the Data & AI Product approach and how to craft impactful Data & AI Products. The last article explained the main causes of failure in Data & AI Products today. This article focuses on the heavy task of defining what is a Data & AI Product.
Happy reading, and we welcome your feedback and any other definitions you have identified as valuable in your experiences.
Every digital product will eventually be powered by Data & AI. That is our conviction.
The key challenge lies in figuring out how to do this effectively while focusing on solving the right problems. It’s not about debating what qualifies as a "Data & AI Product," but rather about embracing a Product Mindset for Data & AI initiatives. The true goal is to leverage Data & AI to elevate and empower products and services.
That being said, clear understanding is crucial to drive alignment and ensure success. This article aims to provide that clarity.
Back to Basics: What is a Product?
When defining a product in the context of software and technology, two standout perspectives offer clarity and depth.
Marty Cagan, in his book "Inspired: How to Create Products Customers Love," offers a comprehensive framework for understanding what makes a product successful. He states:
"A successful product is one that is valuable, usable, and feasible. 'Valuable' means the product is valuable to both the customers and the business. This dual focus ensures that while the product meets customer needs and solves their problems, it also supports the business's goals and is economically viable. 'Usable' indicates that customers can effectively interact with the product, ensuring a good user experience. 'Feasible' ensures that the product can be built with the available technology and resources."
Cagan’s approach emphasizes addressing value risk early, ensuring customer needs and business objectives—like legal compliance, marketing, and financial sustainability—are met in tandem.
Another compelling definition comes from Melissa Perri in her book “Escaping the Build Trap: How Effective Product Management Creates Real Value." She highlights that a product should not be defined merely by its features or outputs but by the value it delivers to customers and the business. As she puts it,
"Products (...) are vehicles of value. They deliver value repeatedly to customers and users, without requiring the company to build something new every time."
Perri warns against falling into the "build trap," where teams prioritize feature quantity over meaningful impact—a concept closely tied to the Data Death Cycle.
Both authors underline the importance of creating products that not only solve real customer problems but also drive sustainable business value.
Connecting the dots between these definitions, we can define a Product as follows:
At its core, a Product delivers a usable experience that creates value for users, aligns with business goals, and is feasible with available technology. The primary focus is on delivering value to the user—when the user benefits, the business benefits too.
In a world defined by unpredictability, embracing uncertainty is key to crafting successful products. A Product-driven approach promotes continuous learning and adaptation, ensuring strategies evolve with the changing environment.
This approach's key advantage is its clear focus on a North Star—a guiding end goal—allowing teams to stay agile and adapt to new challenges as they arise.
So, What is a Data & AI Product ?
DJ Patil first defined the concept of a Data Product in his 2012 book Data Jujitsu: The Art of Turning Data into Product, describing it as "a product that facilitates an end goal through the use of data."
While broad, this definition is both concise and clear, providing a solid starting point. However, despite years of advancements in Data and AI products, confusion persists about what truly defines them. Too often, these products are narrowly viewed, restricting their perceived potential.
Our aim is to offer a broad definition of Data & AI Products, focusing on value creation without being tied to specific technologies or organizational structures. It’s about adopting the right mindset, not just using the right tools.
We use the term Data & AI Product to encompass both Data Products and AI Products, emphasizing their shared characteristics. Further distinctions will be explored in future articles.
This definition encapsulates several key message:
At its core, it's a product. Data and AI are integral but not the sole sources of value. Their true power lies in how they are embedded within a broader product that addresses user needs, blending seamlessly with aspects like design, functionality, and overall user experience.
It encompasses the entire user experience and outcomes. A Data & AI Product must ensure the technology is not only sophisticated but also user-friendly and intuitive. The goal is to enhance the user journey, not complicate it, making sure Data and AI elevate the overall experience.
It’s the result of creative problem-solving. Just like traditional products, a Data & AI Product is developed through creativity, prototyping, and iteration. It must not only solve user problems in a way they love, but also be viable for the business—both desirable for users and sustainable for the company.
What distinguishes Data & AI Products is their core reliance on Data and AI to solve user problems.
By beginning with a Product Mindset, we ensure that the focus remains on delivering value and solving real user challenges, rather than becoming entangled in the complexities of the technology itself.
The ultimate goal is for Data & AI Products to not only harness the power of these technologies but also to meet user needs and drive business success effectively.
What Makes a Data & AI Product So Special?
Data is already a key component in most digital products through KPI tracking, allowing Product Managers to gather live feedback and iterate quickly.
However, digital products that fail to integrate Data and AI seamlessly into the user experience or fully leverage available data will struggle to remain competitive.
The strength of Data & AI Products lies in their ability to process vast datasets, apply AI algorithms, and deliver insights, predictions, and automated decisions. This enables them to solve problems in ways traditional products cannot, offering advanced features like personalization, predictive analytics, and intelligent automation. The use of Data and AI transforms products into dynamic tools that adapt to user behavior, provide real-time recommendations, and continuously improve as they gather more data.
Additionally, Data & AI Products offer deep personalization, tailoring experiences to individual preferences. Most importantly, these products continuously learn from new data, automatically refining algorithms to enhance performance over time.
As Marty Cagan said in its 2024 “AI Product Management” article written with Marily Nika:
“(...) just as how “Mobile PM” was an especially in-demand skill when mobile was new, and now most PM’s are expected to have the skills to develop products for mobile, we expect the same to become true for AI Product Managers. In a few year’s time, we expect most PM’s will need to be skilled at building AI-powered products and services.”
What do you think? Do you see any other valuable definitions of Data & AI Products?
Stay tuned for the next articles to learn how to escape the Data Death Cycle and Craft Impactful Data & AI Products! We will continue to talk a lot about Product Management and its application to Data & AI.
What I like about this article is that it focusses on the value that data analytics and AI brings to the customer. There are many attempts to define “data products” out there and I think many of them miss the mark.
I’m reminded of a small book by Mark Schwartz titled “The Art of Business Value”. At one point he goes a little tautological to say that business values is whatever the business values. It is a joke … but a profound joke. I think of data analytics and AI products in a similar way, the definition of the product is whatever delights the customer. Sometimes this is comprehensive documentation (catalog, lineage, etc) and some times it is in terms of a packaged consumer experience. Product definitions will adapt to circumstance (or YMMV)