AI agents and the future of leadership

2026 was widely seen as a turning point for Artificial Intelligence, as the technology shifted from a niche tool to an essential office companion. ChatGPT, Microsoft Copilot, and Google Gemini all became part of everyday workflows and executive decision-making.

Central to workplace AI’s rapid evolution are AI agents, which are reshaping how leaders interact with technology and influence how their team’s function and the decisions they make.

But what exactly is an AI agent? What risks and opportunities do they present, and will they really give executives the time and space to focus on business growth?

We asked Dr Shaanan Cohney, Senior Lecturer in Cyber Security at the University of Melbourne for his thoughts on whether AI agents are here to stay or a passing trend?

Can you describe what an AI agent is and how they differ from AI chatbots?

Most people experience generative AI through chatbots. They open websites like ChatGPT, Gemini, or Claude, type in a question or prompt, and get a generated response almost instantly.

The catch is that chatbots mainly respond to what you type. On free plans, you can’t usually access all the tools people rely on at work, like spreadsheets, files, company databases, emails, or calendars. Because of this, users often need to manually copy information back and forth between the chatbot and other programs.

AI agents aim to change this. They use same AI models but can also interact with software and tools on your computer. The best example would be downloadable apps like Microsoft Copilot or the desktop version of ChatGPT that can carry out tasks for a period of time and then come back with results, making them more independent and capable than traditional chatbots.

How can AI agents help business leaders work more efficiently and make better decisions?

AI agents have the potential to completely transform how leaders access information and make decisions. Traditionally, senior managers relied on analytics or decision-support teams to analyse data, because they’re busy running a company and they don’t have the time or skills to work through complex spreadsheets themselves. But AI agents can. And Fast.

Now, a manager can upload a spreadsheet, ask a few questions, and get a structured, data-driven response in minutes. In some cases, the agent can even tweak the spreadsheet to do the analysis automatically. That’s where I see real value for senior stakeholders.

We’re also getting to a point where building custom tools doesn’t require coding knowledge. Low-code platforms promised this before, but they were clunky and frustrating. Now, AI agents can create small internal tools that organise reports, summarise information, or automate routine workflows almost instantly.

And this is really just the beginning. The technology will improve. Soon, personal digital assistants will run in the background autonomously, no prompting required, quietly helping you get things done; There are still very big security issues that will need to be worked out, though.

Let’s talk about those security risks. What are the biggest concerns business leaders need to be aware of when it comes to AI agents?

One of the biggest concerns is that these systems can still malfunction, and when they do, the damage to a business can be very real.

At a basic level, once employees have access to AI tools and agents at work, it can be very tempting to rely on them instead of simply using them as a tool. Of course, employees not checking their work carefully isn’t a new problem. That’s partly why supervisors exist in the first place, right?

But the bigger question is this: what happens if relying on AI over time actually weakens our ability to do our jobs, even when we’re paying attention? That’s something that matters whether you’re an employee, a supervisor, or a business leader.

Beyond that, organisations could also expose themselves to cyber-attacks. For example, imagine an employee using an AI agent receives a phishing email targeting the agent instead. It might appear to come from the CEO asking urgently for sensitive financial figures. If the agent believes the request is genuine, it could automatically send the information, potentially compromising the business

This could become an even bigger headache because the temptation to use AI agents is so high.

What can executives do to safeguard themselves and their companies?

Right now, the best thing you can do is give your employees controlled access to internal AI tools so they’re not reliant on their personal Claude subscription to be more efficient at work.

Without that guardrail, the temptation to bypass company rules may well result in a free-for-all where everyone is using AI without any kind of oversight, and well, it’s not hard to imagine the kind of damage that could do.

But the challenge then is that once you officially endorse and rollout AI tools in your organisation the responsibility shifts from the employee to the organisation. You can’t just blame your employee if they misuse the tool because you allowed it in the first place. That’s why having internal policies that mitigate some of those risks is going to be so important.

So, AI Agents are here to stay and not just a passing tech trend?

I think so, yes. So long as we solve the security issues, the value proposition is strong. The big challenge will be how we adopt and integrate these powerful tools at mass scale and in a way that is safe and is in the interests of businesses.

But this technology is advancing rapidly, so it’s really important to make sure that you keep an eye out for developments in this space because tasks that may be inappropriate or inaccessible to an agent today might suddenly be totally within capability two months from now.

Having someone in your organisation whose job it is to keep track of these developments and to actually be pushing innovation forward will give you a competitive advantage because of the productivity benefits they can offer while still protecting your core values.

Featured in this article is Dr Shaanan Cohney

Dr Shaanan Cohney, Senior Lecturer in Cyber Security