The time for AI is now
AI is going to eat the world over the next 10 years.
Just like how traditional software has been eating the world for the last 15 years.
ChatGPT and DALL-E are only the most visible part of the iceberg. AI is seeping into both consumer and business technologies in subtle ways.
For example, a large retailer struggling to retain its sales associates might be exploring if it can use machine learning to predict employees likely to leave. Likewise, a startup of 10 might be interested if it can use machine learning to zero in on the leads that have a higher probability of converting.
For a business to be competitive, the integration of AI into its products, processes, and services will soon enough be as essential as going digital is right now.
If you are a product manager tasked with the adoption of machine learning, it can feel like an overwhelming initiative. This is especially true if you have never led a similar project in the past. It’s very easy to be overwhelmed by complex math equations and other intricacies that are often associated with fields of AI like machine learning.
The good news is that a significant portion of what you do as a product manager for a standard web or mobile product is equally useful when you are developing a product that uses machine learning.
Many principles and practices of product management remain unchanged. Actually, the “customer development” and “ discovery” part of product management is pretty much the same.
As a product manager, your job can be summarized in the following two activities:
Identify the most important pain points or gain creators for the customers you are trying to build a product for
Figure out how you could use design and technology to serve those needs better than existing solutions do.
The core needs of the users remain fairly stable over time. What does change significantly is the technologies you have at your disposal to solve those problems.
ML is just a new piece of technology available to you.
Therefore, the effort you put into understanding your users, capturing their pain points, and understanding the outcomes they value will serve you well in the new world as well.
The solution space is evolving rapidly
Let’s consider how people’s need to listen to music has been addressed over time as new technologies become possible
Gramophones
Cassettes
CDs
MP3 players
Streaming Services
The technology to help people listen to music has changed significantly over the years. The new breakthroughs have allowed innovators to address the pain points associated with listening to music.
If you do your customer interviews well, you can come up with a list of 100+ pain points that get in the way of a user looking to listen to some good music.
Here are a few off-the-cuff examples:
The listeners need to buy and own bulky devices
It’s inconvenient to carry them around.
The selection of songs available is limited.
A big list to choose from means it requires more effort to find the music you like.
As the technology evolved, innovators kept coming up with new ways to make it more convenient for people to listen to music.
And today a lot of these pain points have been significantly reduced. People can listen to music with so much more ease than they did 100 or even 20 years ago.
For example, people generally love the fact that they have a huge collection of music to choose from. It was a hard problem to solve when the music was shipped in cassettes.
The problem was substantially reduced with the arrival of MP3 players like the iPod.
The pain point was almost eradicated with the evolution of web technologies and the subsequent rise of streaming services.
How ML can solve some of the problems better
A huge collection of music to choose from solved one problem but created another.
Consider the following pain point associated with listening to music:
The time and effort needed to find the music you like.
When the only technology you had at your disposal was standard web technologies, you could solve the above problem by creating an intuitive yet powerful search functionality that allowed you to quickly find the music of your choice.
But now you have the power of machine learning.
As you will learn in my future posts, the essence of machine learning is being able to make predictions.
Therefore, you can be more creative now.
What if we could proactively tell users what song they are likely to enjoy instead of asking them to search for something that they might like?
Wouldn’t that significantly improve their user experience?
Yes, it will.
And that’s exactly what machine learning will allow you to do.
How to adopt ML into your product and processes
People still have enough problems associated with many core needs that have been around for a long time - listening to music, watching movies, keeping themselves happy and healthy, becoming productive, traveling, creating content, building products, you name it.
In the realm of business, there is an even bigger array of problems waiting to be solved in sales, marketing, finance, operations, product development, design, people management, etc.
Machine learning suddenly makes it possible to solve those needs in ways not possible in the past.
All you need to do is to figure out how you can take the hammer of machine learning and figure out what nails (problems) you can apply it to.
As a product manager for ML, your job is to figure this out.
In this newsletter, I will help you build the muscle to do precisely that.