AI Accelerator: Sujesh Arukil, Director, Software Engineering

In this installment, we talk to Sujesh Arukil, Director, Software Engineering, for whom AI has moved from a search engine to a real thinking partner. He uses it to pressure-test architecture decisions and to learn new technologies through lessons shaped around the way he thinks, and it’s now part of almost everything he does. Here’s why he believes AI should speed up understanding rather than replace it, and why the human parts of engineering matter more than ever.

The AI Accelerator series is an ongoing collection of conversations with people across Confluence who are at the forefront of how we build and use AI.

From the intelligent features embedded in our solutions to the tools colleagues increasingly leverage to get their work done, each installment hears directly from people across teams and disciplines. Find out what’s working, what isn’t, and what they’re still figuring out. Ground-level perspectives from the people living it every day.

Q: How has your relationship with AI evolved to where it is today?

A: Like most people, I initially used AI tools for simple questions and writing assistance.

I then moved to asking technical questions and realized the AI wasn’t just answering, it was becoming a thinking partner. I could, for example, have an architectural discussion with it, challenge an idea, ask it to argue the opposite side, pressure test decisions, and explore trade-offs in minutes instead of spending hours staring at a whiteboard.

From that point on, AI stopped being a search engine and became part of how I think.

A pivotal project for me was Flex Form Builder, which fundamentally changed how I develop software. Before writing any code, I could load the entire problem into AI – not asking it to “solve” it immediately, but asking it to restate the problem, challenge my assumptions and ask me clarifying questions. Once we agreed on the design, it would produce an implementation plan, I’d review it, reset everything back to a clean state, and then have it build the solution phase by phase with tests after every stage.

AI was changing how I thought. The better I became at explaining my reasoning, context and constraints, the better the conversations became. Eventually it started to feel like I was collaborating with another engineer, which I hadn’t expected.

Q: How are you using AI day to day, across your professional and personal life?

A: Professionally AI has become part of almost everything I do across architecture, documentation, presentations, technical reviews, strategy papers, brainstorming, mentoring, learning new technologies and increasingly software development.

I learn through association, and AI has been hugely beneficial for my learning. Give me a real-world analogy, a mental picture or compare a software architecture problem to something like a school report card or filing taxes, and I’ll remember it forever. Before AI, if I wanted to learn a new subject, I had to find the right book, the right course or the right teacher. Those resources teach one way, to everyone. I was able to guide AI on how I learn, and now every lesson is adapted to me instead of me adapting to the lesson. That’s something no book has ever been able to do.

AI has dramatically reduced the time between having an idea and testing it. I spend less time writing the first draft and more time improving the result. Instead of spending hours researching five different approaches, I can evaluate them in one conversation and focus my energy on making the decision rather than collecting information.

AI has really increased my enjoyment of what I do. Ironically, AI hasn’t made engineering less human. It’s made the human parts of engineering more important. I’ve always enjoyed solving difficult problems more than writing repetitive code. AI removes a lot of the mechanical work, which means I spend more time on architecture, strategy, mentoring and designing better systems.

The only thing I occasionally miss is the satisfaction of building every piece myself. But I’d rather build ten meaningful things than spend my time writing boilerplate.

In my personal life, I use AI for planning travel, cooking, reading history, exploring philosophy and even helping my kids with school.

Q: How does AI help you deliver better outcomes for clients?

A: Ironically, AI isn’t always the answer. First, we must ask what problem the customer is trying to solve. Sometimes AI is the answer; sometimes it isn’t.

But to the question, I see the biggest value in reducing cognitive load. Clients shouldn’t have to understand complex workflows, regulations or data structures just to use software. AI can summarize information, explain recommendations, automate repetitive tasks and help users make better decisions. The technology isn’t the product. Better business outcomes are.

I see the biggest value in reducing cognitive load. Clients shouldn’t have to understand complex workflows, regulations or data structures just to use software. AI can summarize information, explain recommendations, automate repetitive tasks and help users make better decisions. The technology isn’t the product. Better business outcomes are.

Sujesh Arukil, Director, Software Engineering

Q: Were you skeptical of AI to start? Are you still skeptical at all?

A: Like many, I didn’t trust its technical accuracy in the earlier days, but I see much less that concerns me on that front today. I have two primary concerns.

First, I see an unhealthy trend of over-outsourcing to AI. I still write my own thoughts first. I’ll ask AI to challenge them, improve the wording or suggest better approaches, but I want the ideas to be mine. Increasingly, I meet people who let AI generate everything end-to-end. The problem isn’t the quality of the output, it’s that when you ask them why something was done a certain way, they struggle to explain it.

AI should accelerate understanding, not replace it.

The second frustration is context. AI still can’t truly understand an organization’s politics, culture, history and unwritten rules. Those things influence engineering decisions far more than technology does. We still need humans to add that layer into our work.

It is worth noting that AI doesn’t carry the consequences of being wrong. Leadership is still about making decisions with incomplete information, accepting the risk and taking responsibility for the outcome.

Q: Where is AI heading, and what does it mean for the profession and for those starting their careers?

A: Agentic AI is a huge shift, and while I don’t think it’s going to replace engineers, I think it’s going to change how engineering teams operate.

Instead of using one AI assistant, we’re moving toward specialized AI agents that collaborate, writing code, reviewing changes, testing, documenting and monitoring systems.

That feels less like another tool and more like the beginning of a new way of building software.

With this, the value of memorizing syntax is declining; but the value of understanding systems, business problems and communication is increasing. Graduates should absolutely learn AI, but they should also understand why software exists in the first place. As I mentioned earlier, technology is a means to create business capability.

The engineers who thrive won’t be the ones who write the most code, they’ll be the ones who ask the best questions, understand trade-offs and know when AI is helping and when it isn’t.

I also think AI will democratize senior engineering thinking. A junior engineer with AI isn’t suddenly a senior engineer, but they can ask better questions, understand unfamiliar systems faster and become productive much quicker.

For me, AI isn’t about replacing intelligence. It’s about amplifying it. The goal isn’t to think less, it’s to think better, learn faster and spend more time solving the problems that only humans can solve.