Evolving Products into AI Products

In this chapter, we will explore the areas that benefit and negatively impact AI adoption in existing products. Every product will be different and every product owner will come to their own conclusions about what level of AI adoption to infuse into their product. For some products, something as simple as adding one AI feature is enough. For other products, a fundamental change to the underlying logic that powers your product might be required. The decisions of how to transform your product should, first and foremost, be determined by your product strategy and should serve your overarching company vision. These decisions should be collaborative and should come with a high level of executive sponsorship.

As the chapter continues, we will explore various areas of AI transformation, and they will serve as a step-by-step guide to building a product strategy that supports the evolution of your product. We will look at how to approach brainstorming about...

Venn diagram – what’s possible and what’s probable

There are multiple ways in which a product could stand to benefit from AI adoption, and understanding the Venn diagram of what’s probable and what’s possible is an important part of your AI product strategy journey. As you continue this exercise, you’ll go through a spectrum. You’ll start with a really open-ended, right-brain brainstorming session on the cerebral notion of value and how your product can deliver it to your customers, which will then be refined by a left-brain, analytical breakdown of the fruits of that brainstorming session. Both are important parts of the product management creative process.

Whether you look at your product from the perspective of the main problems your customers face, the jobs to be done, or feature parity, you’ll want to get a sense of which AI enhancements would be most high-value for your product, which would be the most low-cost, which...

Data is king – the bloodstream of the company

Before product managers can begin the work needed to start building and developing their product, getting it ready for testing, and releasing it to their customers, they need to get really clear on the strategy of how they will position their product. This is the thought process behind the Venn diagram exercise we saw in the previous section. Now that we’ve gone over the process of how product managers can approach potential AI embellishments, we can add one more layer of scrutiny to this list. This additional layer focuses on the data, which is what will power every single item on these lists. Once product managers begin the work of understanding what they can do with the data sources they have and which data sources they’ll need to make the items on their list a reality, we’re getting close to actually having a plan.

In the following subsections, we will be addressing the key areas of data readiness. Preparing...

Competition – love your enemies

Using the data you already have will take you quite far, but there will always be a need to append this data with a feedback loop from the outside world. Understanding your competitors will help you inform your strategy. It will give you examples from your peers that have already made the jump that you’re looking to make. Some of the examples you see from your competitors will be sources of inspiration and some examples will help you avoid certain mistakes. Doing your due diligence when researching your competitors, particularly those that have also embraced AI, will have some influence on what you choose to build and should be one of the factors that influences the lists we established previously.

Some will refute this point and say that you should build a product based on your own understanding of your market and your customers’ problems. The notion that you should focus on the problem you’re looking to solve without...

Product strategy – building a blueprint that works for everyone

By now, we’ve brainstormed potential ideas. We’ve taken an inventory of our data, as well as the insights that can come from our competition and greater market, and we’re finally able to come to the drawing board and build a product strategy that will reflect the next major era of our product. Going from a traditional software product to an AI product is no small feat and it should be treated as a massive overhaul of the product because so much of how we build, what we build, and how we store, collect, and use data will be majorly renovated.

Building a product strategy will directly influence your product roadmap and this will help you realize which parts of your product will need to transform first to succeed in the actualization of your product’s AI transformation and commercial success. However, there’s a big difference between creating a product strategy that fits in with...

Red flags and green flags – what to look for and watch out for

Moving forward into the wide frontier of AI products includes some common pitfalls in AI transformation, as well as markers for success. We will call them red flags and green flags and they’re signals you can pick up on as you go through this process of getting your product ready for AI adoption. Some of these will be concrete actions or results and some will be more emotional, but either way, you can use them as markers to know whether you’re on the right path or whether you’re hitting potential rough patches. In the following subsections, we’ll address a few red and green flags. Let’s get started.

Red flags

Red flags are behavioral patterns we can try and look out for that indicate there is an issue with a process we’re setting forth. Because AI adoption is such a revolutionary undertaking for any company to take on, it’s best to look out for some of these...

Summary

We’ve now come to the last chapter of the last section of our book: evolving an existing product into an AI product. So much of this chapter has been about how to anticipate and prepare for the jump to AI/ML. In the second section of the book, we heavily discussed concepts related to the AI-native product: a product that’s created with AI initially. Once you do make the jump to fully embracing AI in your own product, you can refer to Part 2 of this book, which is more focused on the aspects that come up when you’re in the flow of building AI. In this section, we wanted to focus on the preparation stages for embracing AI/ML because of the gravity that comes with AI transformation.

Brainstorming ideas, vetting those ideas with practical considerations, getting your data right, evaluating the competitive landscape you’ll be playing in, and bringing in your stakeholders to make a plan for how to build the transition are all part of AI readiness. All...

Additional resources

• Udacity’s Intro to AI (https://www.udacity.com/course/intro-to-artificial-intelligence--cs271) course and Artificial Intelligence Nanodegree Program (https://www.udacity.com/course/artificial-intelligence-nanodegree--nd889)
• Stanford University’s online lectures: Artificial Intelligence: Principles and Techniques: https://stanford-cs221.github.io/spring2022/
• edX’s online AI course, offered through Columbia University: https://www.edx.org/course/artificial-intelligence-ai
• Microsoft’s open source Cognitive Toolkit (previously known as CNTK) to help developers master deep learning algorithms: https://learn.microsoft.com/en-us/cognitive-toolkit/
• Google’s open source TensorFlow software library for machine intelligence: https://www.tensorflow.org/
• AI Resources, an open source code directory from the AI Access Foundation: https://www.airesources.org/
• The Association for the Advancement of Artificial Intelligence...

The rest of the chapter is locked