In our field, we are accustomed to rapid technological evolution and the constant emergence of new tools, which we must learn to work with. However, when it comes to AI, something feels different. In the sense that AI is no longer just a technical tool, but an actor that can influence strategic decisions.
Traditional leadership is based on intuition, experience, and commitment. AI offers speed, data, and apparent objectivity. One of the most challenging leadership dilemmas of the moment is whether AI is going to be an ally that will empower us as leaders – or an adversary that will erode our role and autonomy.
We are faced with a paradox: we can make decisions more easily, with apparently better documentation, but we risk no longer being the ones who decide.
What data says
Data shows that AI is now commonly used across the software development lifecycle. For example, DevOps research shows almost all organization’s IT leaders use AI to improve security and code quality: 64% use AI for security assessments, 60% for automated code testing, 59% for detecting breach patterns, and 58% for vulnerability scanning. Other popular uses include automating repetitive tasks (40%), generating UI layouts (34%), and detecting bugs (32%). However, only ~30% of teams report that automation truly frees developers for strategic work.
The role of leaders in AI adoption
Strong leadership is often seen as the deciding factor in whether AI adoption succeeds or fails. Experts agree that introducing AI isn’t just about the technology – it’s about people and leadership. Success usually depends on things like having visible support from executives, aligning AI efforts with business goals, fostering the right company culture, providing training, and empowering grassroots champions within the organization.
For over a year now, I have been trying to gauge the best leadership methods and also, the biggest challenges that we face at RomSoft. We try to organize meetups at least once a month to discuss these ideas. The topic of AI-assisted leadership has naturally crept onto our agenda. I have tried to extract the biggest concerns we have and see how they overlap with the concerns of other leaders around the world. Here are some general concerns, but also some specifics:
General concerns about AI of leaders world-wide
🟢 Overreliance
This concern tries to address the risk that important decisions will be left entirely to AI, giving up intuition, human context, or experience. Or even decisions that may be “pre,” “post,” or “coordinated” by AI without sufficient human oversight or verification.
In this report, Goldman Sachs warns that they have identified a “risk of over-reliance”.
🟢 Transparency and traceability
This concern refers to a reality that we confront every day: AI tools often do not clearly explain how they arrived at a particular recommendation, what data they used, what were the decision criteria. How can we, as leaders, without a context, assess if the recommendation fits our organizational environment, or if, on the contrary, it may lead to unexpected effects?
🟢 Bias, data errors, or bad data
Another problem that we are aware of is that AI can reflect and even amplify existing biases in data. If training data is not representative, decisions may disadvantage certain groups or produce unintended consequences. Here’s an article about how leaders can positively address bias in AI decision making.
🟢 Responsibility
This is a major concern that ringed true to everybody in our group – If an AI based decision leads to negative impact (financial, reputational, legal) who is responsible? Who takes the blame/ supports consequences? Is the responsibility split between the organization leader/the tool developers? Is it even possible? AI accountability is widely recognized as a critical global issue, essential for building trust and ensuring sustainable adoption.
Other concerns may include team trust and acceptance, data privacy, hidden integration costs, loss of human nuances and lack of diversity in decision making (risk of herd behavior).
Particular concerns of team leaders in RomSoft
In our latest internal survey, there are some particular concerns that I could identify among our team leaders:
🟢 People will complete tasks based on incomplete data
AI systems are only as good as the data they’re trained on. If the dataset is incomplete, biased, or contextually shallow, the AI may provide outputs that look convincing but lack depth or accuracy.
Leaders worry employees may trust the AI output too readily, skipping the critical validation step and making decisions on flawed or partial information. For example, over-reliance on dashboards or predictive models without understanding their blind spots.
🟢 Cost of output control is higher than anticipated
AI is often promoted as a cost-cutting tool, but hidden expenses come from the people and processes needed to oversee its work (supervision), check its accuracy (validation), and ensure it follows rules and policies (governance). These can sometimes outweigh the expected productivity savings.
🟢 Adoption just based on FOMO, in attempt to ride the hype wave, prematurely giving up important skills that are already acquired
Many organizations feel the pressure to implement AI because competitors are doing it or because of hype in the media. Leaders fear premature adoption without strategy, leading to loss of investment if the tool is abandoned, skills atrophy (employees stop practicing critical abilities because AI is assumed to handle them), or lock-in with immature vendors or tech that won’t scale.
To make a real world parallel, we all remember past hype cycles like blockchain-for-everything or VR in corporate training – projects launched quickly, then abandoned, with opportunity costs and weakened internal expertise.
Key Takeaways
AI is transforming work within software development companies, but its success depends on thoughtful leadership and strategy. While tools like generative coding assistants and automated workflows are very powerful, executives must lead the change by setting vision, fostering collaboration, and building an AI-human collaborative culture.
When companies pair strong executive support with ideas driven from the ground up, they unlock major gains in both productivity and innovation. But if they overlook the people and processes behind the technology, even the best investments can lose momentum.
We can call this approach holistic, as it aligns AI use with business goals and team culture. By empowering engineers, rewarding experimentation, and maintaining ethical standards, business leaders can use AI not just as a set of features, nor as a standalone player in decision making, but as a strategic advantage for innovation and growth.