AI and the future of product engineering

In this post, I'd like to explore the role of the software engineer in the world of product development, how it is impacted by AI, and what we can expect from this transformation.

TL;DR

• AI enables engineers to spend less time on mundane tasks, freeing up their time for tasks AI cannot automate yet
• AI-enabled engineers with product experience—similar to indie hackers—take part in the whole product development lifecycle
• Hiring product engineers is a resource-efficient way for companies—especially startups—to build products users want

Engineering and AI

I don't think AI is going to replace software engineers. But since it enables us to automate many chores, we can dedicate more time to tasks that matter, such as solving real user pain points.

The shift from purely writing software to spending more time on the product and its users already gave rise to a role that's gaining a lot of traction: the product engineer.

What exactly is a product engineer?

Let me quote the definition provided by the team at Posthog:

A product engineer, at its most basic, is a software engineer building products. They do similar work to software engineers: writing code and shipping features. [...] What makes them unique is their focus on creating a product for users.

I would even go as far as saying that product engineers are the indie hackers of the modern startup. They don't only know how to deliver software but also care deeply about validating it against the market.

Let's take a look at Google Trends to see how the term "product engineer" has evolved over time.

As we can see, that's a steady growth curve with a burst in 2022 (probably thanks to Posthog manifesting what a product engineer does). And I think this role becomes even more exciting with the rise of AI. Let's explore why!

AI makes coding cheaper

AI can write pretty good code and returns complex algorithms based on prompts that take a fraction of the time to write.

Of course, the output often contains bugs, and debugging them is a lot easier with a strong software development background. AI assistants like Copilot and Cursor AI obviously speed this process up significantly.

Also, AI development could reinforce biases and automate bad practices if not closely monitored. But let's imagine a world where all of this has been perfected to a point where you rarely have bugs and minimal bias.

Most programmers should gain a lot of time, time they can spend on high-level architecture, code reviews and fixing bugs.

The worst case—the dystopia we all fear—is that engineers become obsolete. In an ideal world though, aren't we now simply able to ship software faster?

Maybe, but speed doesn't automatically match better outcomes. You can write new features in less time, but someone still needs to make sure that the product is aligned with users' actual needs and solves meaningful problems.

Say hi to the Product Engineer

With their technical expertise and focus on the user the product engineer is well positioned to benefit from this transformation.

If writing code takes less time, the product engineer can spend more time on other stages of the product development lifecycle, such as user discovery, product strategy and analytics, thus becoming a more holistic product expert.

Let me dig deeper into what I think the responsibilities of the AI-enabled product engineer of the future might be:

Product Discovery

Most product development starts at this stage: analysing your users, gathering insights, and understanding the market and your competitors.

As of now, this doesn't seem like a task that AI can do autonomously, even though it can aid in the process. It requires:

• Empathy
• Real conversations
• Deep insights analysis

Activities at this stage usually involve lots of calls, whiteboarding, brainstorming sessions and stakeholder meetings.

Tools like Fireflies.ai are great at accurately transcribing calls and running prompts on the results. I think they're becoming an indispensable tool in product discovery.

But interpreting those insights, eliminating bias and errors, and identifying the signal among the noise is a task that requires human reasoning, experience and intuition.

Product Strategy

While I wouldn't expect developers to be experts in business strategy, there are enough strategic decisions in the product lifecycle that an engineer has to take on a daily basis.

It includes:

• Backlog prioritisation
• Defining roadmaps
• Aligning with business

Doing this requires some degree of business acumen and a lot of product experience. It involves planning, anticipation, and continually re-evaluating priorities.

While LLM-based tools—like Squad—can help automate the grunt work and speed up the process, they don't fully replace the complex decision-making process. That's where we as humans are still at an advantage because we can feel how certain decisions play out in the future and envision their consequences in the real world.

Engineers with deep domain knowledge are obviously in a much better position to use their judgment and make good strategic product decisions.

Product Delivery

Last but not least, I want to cover some technical tasks that I think AI is still struggling with as well:

• System design and architecture
• Migrations and refactoring
• Bugs and chores

They need you to have a deep understanding of the codebase and its nuances, but also an understanding of the product objectives—tasks that involve reasoning, spatial thinking, and a lot of nitty-gritty configuration work.

AI is great at supporting us in this process. Tools like Qodo claim to help developers write, test and review code.

But it is probably too soon for an AI to autonomously plan, design and deliver bug-free user experiences, or refactor and migrate large parts of a codebase, especially when a system is highly distributed and complex.

A glimpse into the future

Product engineers are well-positioned to tackle each of these areas. They understand both, the technical implications of product decisions, and can emphasise with users. These skills and the ability to employ AI at each of these steps make them an invaluable asset for any product team aiming to be more resource-efficient.

This doesn't mean non-technical people can't utilise AI to the same extent. Also, UX designers, business analysts, and product managers bring a depth of expertise that engineers, even AI-empowered ones, might not easily acquire. Most often, it's the discussions between people from diverse roles that lead to the most interesting results.

However, engineers can use no-code tools more efficiently and build prototypes quickly. Combined with an eye for good design and an intuition for what users want, this is a skill which can accelerate product development significantly. My experience is that people working at the intersection of product and engineering are indispensable team members, and they're often the glue between business and code.

Don't get me wrong. I don't expect all engineers to become product experts. Despite the advances of AI, there's still a ton of deep technical expertise required for highly technical roles.

From a business perspective though, it's an exciting opportunity: product engineers can contribute at all levels of the product development lifecycle, potentially resulting in less communication overhead and fewer misunderstandings. The upsides are increased productivity, better alignment on user goals and—most importantly—increased user satisfaction.

If you would like help in understanding the role of the product engineer in more detail, feel free to get in touch via LinkedIn.