What's it really like to be an engineering leader in 2025? We have the answers.

We built Jellyfish around the idea that the best engineering and business decisions are based on data. Relying on gut instinct and assumptions never leads to the best result.

But when it comes to the engineering experience, numbers don't always tell the full story. The metrics may show most engineers are using AI, but how do they feel about its impact? Why are some people still reluctant to adopt AI? What challenges are developers facing in 2025, and what are their predictions for the next few years?

We asked more than 600 engineering professionals to share their experiences for this year's State of Engineering Management report. The responses suggest both uncertainty and optimism among software engineering teams and their leaders. 

Here are the key takeaways.

Change is a constant challenge 

Engineering leaders are there to guide teams through difficult times, but a lot is beyond their control. Shrinking budgets and global economic challenges are enough to test even the most effective leaders and their teams. For smaller organizations, that uncertainty can be particularly tricky to manage. 

“The economy is likely to test my team more aggressively than they have been tested before,” said an engineering leader at one small company. 

In every team, some people adapt easily to change while others struggle to keep pace. For one UK-based engineering leader, repeated periods of instability highlighted these differences. 

“I've led the organisation since Covid times, mitigated the effects of departure of previous VPE, mitigated significant cost-cuts, removed several non-performing managers, directors and tens of engineers, managed several re-organisations. I see around 25% of my staff being able to drive the change, around 35-40% are able to cope with the change and the rest usually struggles.”

To protect the engineering organization, leaders need to do everything they can to build agile, resilient teams. That means providing engineers with the tools and resources they need to stay effective and productive, regardless of what’s going on outside the org. Running a DevEx survey is a great starting point for leaders trying to figure out the best way to support their teams. 

Strong opinions on AI

While AI coding tools can take repetitive tasks off engineers' plates, they also add pressure. According to one respondent: 

"With the rise of LLMs, engineers are expected to be more productive than ever to stay relevant."

But producing more only makes sense if engineers know what they're working towards. “There is this ‘move fast’ attitude without a corresponding goal/outcome to go after; it's not pivoting, it's chaos,” added another participant.  

Many engineers are starting from scratch with AI, learning how to use a variety of tools and verify AI-generated code. This could demand even more from engineers, at least in the short term. 

"There's learning to be done as well as the planned deliverables, so people may have to make more effort," said one respondent. 

Another leader pointed out that getting up to speed with AI isn't an easy fix: "There is a skills gap among existing developers, and training them requires additional time and cost."

But integrating AI into engineering workflows also brings benefits. With AI, engineering teams can focus on the high-impact work that offers the greatest job satisfaction. Instead of getting bogged down in routine tasks, engineers have the time and space to develop creative solutions and contribute directly to the company's objectives.

For some organizations, that shift is already having tangible results. “Over the past six months, we have observed an increase in strategic roadmap work vs. KTLO, which coincides with the period that AI tooling has been adopted,” said one engineering leader.

Making sense of increased AI adoption

According to the 2025 State of Engineering Management report, 90% of teams are using AI, up from 61% last year. 

The data also shows engineers and leaders are mostly positive about AI and how it fits into their workflows — six in 10 respondents believe AI coding tools improve developer velocity and productivity by at least 25%. 

With AI in the spotlight, it's easy to assume coding tools can do anything engineers ask for. But even AI has its limits, and that's holding some engineers back: 

“I'm concerned that the hype doesn't match current capabilities, leading to difficulties in adoption among engineers,” said one respondent.

Other professionals are worried about the quality of AI-generated code. “Junior/less skilled engineers/people with no software engineering background can generate 10x more buggier/low performing code (or in general, code that is not aligned with IDP) thanks to AI," said one engineering leader. 

A different engineering manager shared this sentiment, "In my point of view, AI will increase the velocity of code writing, but it will not necessarily be better code. In organizations where the Dev is not very senior – it might, but in complicated use cases it might not perform as well as a senior Dev or Architect.” 

Engineering leaders need to understand the impact of AI before it can transform their organization. With Jellyfish, leaders can see which tools perform best across different tasks and help engineers use them in the right way.

AI is a companion, not a competitor

The impact of AI on software engineering is only just beginning, and AI coding tools are changing rapidly to deliver more value to users. But no matter how those tools evolve, engineering leaders are certain of one thing: AI cannot replace human engineers. AI can help engineers enormously when it comes to generating code and fixing bugs, lifting some of the burden off busy developers. But AI lacks the human vision and understanding that makes skilled engineers so valuable to high-performing teams. AI is also empowering employees outside the engineering organization to experiment with “vibe coding” — which prompted nuanced opinions from our survey respondents.

According to a product ops leader:

"AI is awesome, but too many take it at face value and that currently is a huge problem. While AI can help creatives, AI itself is not creative. If you have smart people using AI who also understand the topic/issue they are going after, magic happens. Otherwise, you have people that desperately just want to look like they have done something amazing, but don't really understand the issues they have just created with the help of AI. Smart, knowledgeable people are still essential to make AI useful.” 

For more observations from engineering leaders, as well as data and insights from our 600+ global respondents, download the 2025 State of Engineering Management report.