Engineering Teams Accelerate with AI: Navigating the New Landscape

Understanding What to Build in the Age of AI

Engineering teams are accelerating thanks to AI. But where’s the bottleneck now? It's knowing what to build.

I don’t usually make predictions, but here’s one: the skill of understanding and expressing customer needs is going to become way more important. If I'm wrong, I promise to eat my words and explore why.

AI is Blowing Up the Old Delivery Model

AI is reshaping software development, and it’s not just about speed. Here are some key points to consider:

• 82% of engineers are using AI to write code. Check it out here.

• Velocity is up 30–50%. However, DevOps, QA, and security teams are under pressure.

• Tools like ChatGPT are turning engineers into editors instead of authors.

  
  

This means teams can build more, faster, but only if they know what to build. Right now, most don’t.

I’ve seen a lot of vibecoding, and while those projects ticked the boxes of the assignment, they often miss the mark. They accomplish their goals, but they aren’t designed for customers. Many of these projects simply aren’t good. At my startup, we use AI to explore ideas, but we always seek feedback afterward.

To create something that truly delivers, we need to understand the customer and the job they want to accomplish. This understanding is crucial for effectively communicating with AI.

Research is Coming Back, But It Must Be Different

As engineering output accelerates, we risk hitting a wall of bad bets. We could end up with features that don’t land, tools that don’t help, and products that solve non-existent problems.

That’s why the skill of UX research is making a comeback. It’s not just about the job title; it’s about the core muscle: understanding people, their context, their goals, and what “good” looks like to them.

Good data is hard to find on tech layoffs by role, but research was laid off at a higher rate than designers. I’ve seen entire research teams eliminated. There are many reasons for this:

• An established backlog of work.

• Economic pressure to reduce headcount, paired with a belief that designers or PMs could conduct their own research.

• Discontent with the return on investment. I’ve heard a lot of executive frustration with teams, although I haven’t seen data on this.

• Outsourcing research is easier than outsourcing design since researchers don’t need to collaborate as closely with engineering, product, or the design system daily.

  
  

We face new problems and customer expectations. Our leadership operates in an economic environment that demands doing more with less. We can’t work the old way. We need to leverage our skills and empathy to find new ways of working and delivering value.

How We Need to Change

To adapt to this new landscape, consider the following:

• Your AI-enabled team can build much faster, so your upstream strategy, insights, and prioritization need to scale too.

• Generative AI produces multiple right answers, and the results will vary. We need to provide direction on what a good result looks like and establish guardrails on what to avoid.

• Adopting these new tools presents challenges and requires understanding and scaffolding to ensure success.

• Organizations need new methods for governing these tools and managing risk. Research can support these efforts with real data and understanding.

  
  

What This Means for You

If you’re in product, design, or strategy, here’s what I recommend:

• Reinvest in discovery. Talk to customers. Validate assumptions. Don’t wait for post-launch analytics.

• Redesign the research role. Focus less on lab time and more on fast cycles, integration with go-to-market strategies, and helping prioritize the right bets.

• Hire for understanding, not just speed. Empathy, critical thinking, and storytelling are essential.

  
  

Understanding people and guiding decisions is always necessary. However, this won’t be a return to the research team structure of 2018. We’ll see new job titles, tools, and practices emerge.

The old version wasn’t cutting it, and even if it was, the needs have changed.

If you’ve noticed the same shift, I’d love to hear from you. Let’s figure it out together.