Human In the Loop: Five Project Managers on Navigating AI

Viktoriya Zinovyeva, Project Manager

Almaty, Kazakhstan

I see people burning out and taking it very personally and emotionally. “When should I learn all this new stuff? How should I prepare myself for this new life?” This is common to everyone in IT, not only project managers.

I think that we will see production teams where we have just three people, not 10. And these three people will produce the same amount of work, with the same quality, maybe even faster. Team size is already shrinking. But as a PM, you need to learn how to use AI to fill these gaps, and to respond to these expectations.

Project managers will be expected to shift either to the technical side, being more like an architect, potentially, if they have technical skills, or to the business side, being more like a product owner or product manager. Because nobody's going to pay for so many different roles; you have AI, after all.

There will be a new job for project managers – orchestrating hybrid teams of AI agents and humans by setting up new processes for mutual (symbiotic or centaur) teamwork. They will have to set up new environments automating the routine (with AI tools, skills, proper context management) while leveraging humans to do oversight, judgement, and strategy work.

What we can say for sure is that if you are the kind of PM whose default mode of work is being an administrator, like moving tasks in Jira and pinging your developers, then definitely AI will replace you.

I think it's only a matter of time before my work is questioned. And the question is, what is your new added value in this new environment? What value do you add to the project that AI cannot give?

Dmitry Bulavin, Project Manager

Kharkiv, Ukraine

Being an expert PM today means moving beyond delivery and becoming more of a strategic partner to a client. It's no longer enough to manage a Jira board, or place a task from one column to another column. We must understand the underlying data architecture and the business problem which we need to solve.

For example, we need to have techno-commercial intuition. AI doesn’t have this. We need to see how we can start from a POC for $65,000, and transform this solution into a $360,000 enterprise transformation. We need to be not only the manager who works with the team and listens to the client, but also the strategic partner who can understand the solution, and understand what the client wants as a business.

AI is faster, of course, than any person. But we have this intuition that can’t be produced by artificial intelligence. I think we need to work with artificial intelligence more on routine tasks.

Elena Verbitskaya, Delivery Manager

Odesa, Ukraine

Despite all the hype, I think AI is not as intelligent as people claim. It's just algorithms—sophisticated ones, but fundamentally algorithmic. It will never be as intelligent as humans. I have a PhD in artificial intelligence and know the math behind it.

In the nearest future, we will not have differentiation by software engineer skills like Python or .NET anymore. These will become redundant. The industry will shift toward valuing other competencies—architectural thinking, problem-solving, domain expertise. The basic coding skills won't be the differentiator anymore.

In one of my previous accounts, we experimented with implementing an AI-powered development workflow that resembled a dark factory setup. The process was quite remarkable: an engineer would start with a call with their manager, and an AI-generated transcript would automatically create a Jira ticket. Another agent would take the ticket description and generate implementation using predefined skills or libraries. Additional agents handled testing, verification, approvals, merging, and deployment to production. It was almost completely automated.

What we learned from that experience was that this approach works exceptionally well for startups without big client databases. However, with legacy projects and big clients, the equation changes entirely. You need to check every step the agent produces. The risks are completely different depending on your company and situation, and that particular implementation taught us the importance of contextual decision-making.

Tokens cost a lot. On certain projects, people are actually less expensive than using AI infrastructure. So it really depends on your goals and what you're trying to achieve.

Karen Grigoryan, Project Manager

Berlin, Germany

Being an expert project manager today is less about following a strict process and more about making a significant impact, adapting to changing circumstances, and making informed decisions. The role has evolved from solely focusing on delivering projects on time, within budget, and meeting specific requirements, to driving outcomes and generating business value.

In the face of rapid technological advancements, project managers must possess the ability to work with limited information, make quick decisions, and iterate swiftly. Simultaneously, they need to adopt a more business-oriented mindset and prioritize outcomes.

Technical proficiency is also becoming increasingly crucial. While project managers don’t necessarily need to write code, they should have a solid understanding of APIs, system architecture, AI capabilities, cloud technologies, and data flows. This knowledge empowers them to make well-informed decisions and effectively communicate with engineering teams.

One aspect that remains unchanged is the role of project managers as connectors between various stakeholders, including business, engineering, customers, and other relevant parties. However, today’s role demands a broader knowledge base and enhanced strategic thinking skills.

Overall, the PM profession is not disappearing – it’s evolving. And those who embrace change, learn continuously, and adopt new technologies will thrive.

Andrey Sadakov, Delivery Manager

Novi Sad, Serbia

Honestly, I find what AI is doing to our industry absolutely thrilling. After years of managing teams and thinking about software from the outside, I've started building things myself with AI. That shift alone tells you something profound is happening.

The pace of change is staggering. New model improvements land almost every week, and keeping up isn't optional – it's the job now. But here's what I think most people get wrong: they measure AI's impact at the individual level. The real performance multiplier happens at the team and company level, when you redesign the entire flow rather than just automating one step in the middle.

That demands a different kind of thinking. Systemic thinking. Not "how do I go faster here?" but "how does value move from idea to delivery, and where does AI reshape that whole journey?"

Critical thinking matters just as much. AI can make you sharper – or it can quietly make you dependent and shallow. The difference lies in how you use it. Treat it as a vending machine for ready-made answers, and you'll gradually stop thinking for yourself. Treat it as a thought partner for brainstorming and pressure-testing ideas, and the results can be remarkable.

To thrive in this era, I'd bet on three capabilities: systems thinking, a genuine product mindset, and solid architecture instincts. These are the lenses that let you direct AI purposefully – rather than just surf the wave and hope for the best.