AI Will Transform The Products That We Build – Here’s How Product Managers Can Prepare
Recent advances in artificial intelligence will change how we build, use, and interact with technology. Here’s seven things that product managers can do to prepare for an AI-powered economy.
“Artificial intelligence” is a buzzword that is thrown around a lot in technology circles, often to describe products that are not very intelligent at all.¹ However, AI has come a long way and is set to revolutionise many aspects of our lives. In particular, recent advances in generative artificial intelligence (Gen-AI) are set to disrupt how we use and interact with technology in our daily lives. In this article I hope to show you why it is reasonable to believe that AI will change your job as a product manager over the next decade, and why data science literacy should become a part of your toolkit.
The state of AI in today’s tech landscape
Artificial intelligence (AI) is the ability of a computer to exhibit human-like behaviour such as learning, problem-solving, and decision-making. In other words, AI allows a machine to perform tasks that normally require human intelligence.
To be sure, AI has had a rocky history in the real world, as researchers have tried and failed to build machines capable of sophisticated thinking. Recent advances in machine learning and artificial neural networks, however, have changed the game. Computer systems can now do things that humans previously could only do. It’s time we take notice.
Generative articificial intelligence has the potential to transform how we build and interact with technology.
Generative AI (Gen-AI) in particular has the potential to transform how we interact with or even build technology. Gen-AI can be defined as artificial intelligence that can create new content, such as music, essays, or images. These new systems, built on deep learning models trained on massive datasets, have huge potential because they can be applied to different contexts with little to no fine-tuning. As a consequence, AI tools take less time to develop and even people without specialised skills or technical background will be able to integrate AI into their products.
Coupled with the advent of the Low-Code and No-Code movements, generative AI has the potential to speed up the transition to a truly AI-powered economy, where AI applications become an integral part of our daily and professional lives.
OpenAI’s launch of ChatGPT has put Gen-AI firmly on the map. But, ChatGPT is far from the only tool that should capture our attention — a whole new wave of tools and applications is emerging that is built on state-of-the art AI. Let’s look at a few examples.
GitHub Copilot uses AI to auto-generate code from text prompts. It relies on the OpenAI Codex AI, and auto-completes your code, or even suggests complete function bodies based on what you type.
In turn, a GPT-3-based model called DALL-E, and its next-generation cousin DALL-E 2, can generate photo-realistic images from text prompts, opening up a whole new range of applications in the creative space.
Imagine an interior designer who can ask a machine to generate a photo-realistic image of a room, or an advertiser who can quickly generate images for a campaign. Or, a UX designer who can quickly generate wireframes or even high-fidelity designs through a simple text prompt. The list goes on.
You can try out DALL-E 2 here.
Further, the field of natural language processing (NLP) has evolved at lightning speed over the past couple of years, to the extent that some models now have the potential to add real value for customers.
For example, in content marketing and SEO optimisation, products have emerged that use NLP to help writers co-create high-quality content. These include the likes of Mark Copy AI, Writesonic, and Jasper, Rytr, Wordtune, and Writer, to name but a few examples.
If you have experience as a writer, you’ll know the feeling of the much-feared writer’s block. What if you could simply push a button, and have a model generate a couple of lines for you to help you along? Sounds crazy right? Yet, these tools are available today, and they’re quite impressive.
What does it mean to be a writer in a world where AI-generated writing is indistinguishable from human-generated writing?
These technologies have been around for several years now, but I am confident that we’ll see some of the most profound advances within this decade. Beyond improvements in the models themselves, we will see advances in their application to real-life problems, helping humans in their day-to-day jobs. And these advances will transform the way we work.
For example, what does it mean to be a writer in a world where AI-generated writing is indistinguishable from human-generated writing? What does it mean to be a musician in a world where AI can compose music at scale? And, what does it mean to be an interior designer in a world where AI is capable of producing genuinely creative designs? These are questions that we’ll start to see answered over the next few years.
An AI-enabled future
These opportunities will come with challenges too. How will humans interact with these tools? Who owns the data that is generated by these models? And, how will humans cooperate in a world where AI is capable of doing things that we’ve never seen before?
If you want to be ready for the next five to ten years, it’s important to understand what will be possible in this time. Below is a quick list of things that I believe will happen.
We will see a proliferation of embedded AI applications — AI will become more widespread, showing up in places like vehicles, home appliances, and tools. These applications will be more accessible than ever before, and we’ll start to see them used for a wide range of purposes.
We will see further advances in natural language understanding. These advances will make it easier for people to interact with their devices, and will also open the doors for new applications that we haven’t even envisioned yet.
We will see the growth of human-AI teams that collaborate both with each other, and with AI-enabled software and tools.
Within the next five to ten years, we will see the rise of a true AI-powered economy.
We will see the rise of a true AI-powered economy as companies emerge to solve problems that are currently unsolved. AI models will unlock new opportunities, and we’ll see a wave of startups emerge to tackle these opportunities.
We will see a rethinking of our relationship with machines. The same way that AI is making our lives easier, it’s also making it possible for machines to do things that were previously unimaginable. As a result, the ways in which we think about and interact with machines will change dramatically.
We will see an increase in creativity among humans, as AI makes it possible to do things that humans can’t do on their own, or are less well equipped to do on their own.
Seven strategies for product managers that want to be ready for an AI-enabled future
So what will it take for product managers to be prepared for this kind of future, and to harness its opportunities? In my view, the solution is for data science literacy to become a part of every product manager’s toolkit.³ A foundational understanding of some of the principles of data science is required for all PMs, given how pervasive AI is going to be in the near future. So, here’s a list of practical things that PMs can do to get up to speed.
1. Stay up to date on the latest and greatest in AI
First, make sure you stay up to date with what’s happening in the AI space, and how it is making its mark on society (and your customers). The following blogs are a great place to start:
Books are also be a good way to get into the fundamentals of AI. Here is a list that may help you get started:
- Factfulness: Ten Reasons We’re Wrong About the World — and Why Things Are Better Than You Think (2018), by Hans Rosling, Anna Rosling Rönnlund, and Ola Rosling
- Algorithms to Live by: The Computer Science of Human Decisions (2016), by Brian Christian and Tom Griffiths
- Naked Statistics: Stripping the Dread from the Data (2014), by Charles Wheelan
- Weapons of Math Destruction (2016), by Cathy O’Neil
Or, if you’d prefer to read a summary and review of these books, check out this blog post 👇
2. Take courses
Second, take courses that focus on the concepts and fundamentals of AI. DeepLearning.ai’s “AI for everyone” course is a great place to start.
3. Keep track of AI-based software and applications
Third, try to keep track of what AI-enabled products are emerging in your space. Check out this overview of apps and companies that are already using GPT-3 (a neural network with 175 billion ML parameters parameters!) to bring the power of AI to their customers. In turn, futurepedia provides an overview of AI tools (569 and counting) across many categories, including gaming, writing, and image generation, to name but a few.
4. Know your company’s data … and how it’s used
Fourth, take stock of the data that exists in your company. What is its origin? What are the challenges of wrangling and storing this data? What is its information quality? Can the data add value for the customer, and in what ways? Sit down with your data science team and your data engineers, and try to dig into the details.
5. Up your GPDR game
Fifth, familiarise yourself with the regulatory challenges that come with processing and storing data. As a product manager, you’ll need to be attuned with changes in the regulatory environment and how it affects your products (e.g.: GDPR, the California Consumer Privacy Act, CCPA). If DPAs and DPIA are abbreviations that do not yet sound familiar to you, make sure you learn what they are for (even though they may not be your direct responsibility). (On a side-note, while I say “regulatory challenges”, I should add that GDPR is actually one of the best things that has happened to the tech industry in a long time — but that’s a topic for another day).
6. Be aware of the pitfalls of AI
Sixth, as a PM you need to be constantly aware of the limitations of our data and how that may affect your customers. Any biases in our data will also affect our customers (yes, I am going to go ahead and use the “with great power comes great responsibility” quote here). Sometimes, we are only made painfully aware of our products’ negative impact once it is already too late and out there in the market. (Cathy O’Neil’s Weapons of Math Destruction is a fantastic read on this topic, and a powerful lesson on how poorly designed algorithms (or abuse of algorithms) can have real-life adverse consequences.)
7. Keep an open mind, and be creative
Seventh and finally, we need to open our minds to completely new opportunities, and set different expectations of how our customers will interact with our products in the future. Software often acts as an (inefficient) interface to translate what is in our minds to a digital form. AI has the potential to reduce many of the frictions that hamper the creative process.
The very notion of a “computer” will be broadened beyond the traditional keyboard and mouse input and output. How do we create better software that takes advantage of these new opportunities? That’s a question we should always keep front and centre as we go about our daily work as product managers.
The future is exciting. Now let’s build it!
P.s. parts of this article were co-written with an AI tool. How’s that for an AI-enabled future? 😉
Notes
¹ In this article, I’ll use the term “artificial intelligence” rather liberally, in the general sense. Artificial intelligence refers to the ability of a machine to exhibit human-like behaviour such as learning, problem-solving, and decision-making. AI allows a machine to perform tasks that normally require human intelligence. Machine learning is a method that allows a computer to learn from data, solving tasks without being explicitly told how. In turn, deep learning is part of the machine learning family, and employs neural networks with multiple layers. Many of the examples in this blog post are actually applications of machine learning, deep learning or of natural language processing (NLP).
² I deliberately use the term “data science literacy” here (rather than simply “data literacy”), because the role of a data scientist typically encapsulates a wider range of skills that are relevant to understanding and productionalising AI, including, for example, machine learning (and possibly deep learning and natural language processing too), some data engineering, developer skills, statistics, and mathematics.
If you liked this article, you might also enjoy these blog posts👇
Please read this disclaimer carefully before relying on any of the content in my articles on Medium.com for your own work :)
