Technology is a fantastic thing – it’s hard to believe that in a little over a century, mankind has gone from Marconi’s invention of the wireless telegraph (also known as radio) to the rapid proliferation of modern computers and the Internet.

As technology has evolved, its underlying capabilities have also dramatically shifted. Gone are the days of simply command-line inputs on computers – instead, our devices can do an incredible number of things, even in form factors no larger than the palm of your hand.

In recent years, software that has leveraged emerging trends in artificial intelligence has begun to preview a world where some computers are developed with an element of autonomy, allowing them to make some decisions independently. This innovation is known as agentic AI and represents a new frontier in what’s possible with artificial intelligence.

What does this mean for this generation of business graduates? After all, what is an MBA if the decision-makers are all machines? As we’ll soon discover, combining traditional qualifications such as an MBA with new and emerging technologies like agentic AI can help business professionals stand out amongst their colleagues.

AI – a rapidly evolving frontier

Artificial intelligence may seem like a field that has only recently been discovered. However, you may be surprised to learn that the idea of automated processes dates back centuries, to the days of Ancient Greece and automatons.

AI has rapidly evolved in the last forty years. It is no longer just a high-level film concept like James Cameron’s sci-fi thriller The Terminator. Instead, AI has begun to embed itself in every part of our lives, whether in content creation or providing results and information.

The concept of AI is remarkably straightforward to understand – take any process and automate it through technology in such a way that it can make decisions independently and unaided by human operators. While the concept seems remarkably simple, as we’ll soon find out, one of the challenges of such independent thinking is that numerous random and unforeseen obstacles can make a problem much more complicated than first envisaged.

The complexities of automated decision-making

In the early days of technology, automated decision-making seemed like it would be reasonably easy to solve – ideating on what’s possible through the use of process automation. As technology evolved in fields such as computer vision, new process automation methods, including RPA (robotic process automation), became mainstream.

While process automation can be straightforward to understand, it does have limitations. For example, if system requirements are constantly changing, RPA may simply be unable to adapt without help – requiring updates and changes to deal with an increasing multitude of use cases.

Even when humans are involved, process automation can fail. Many customer support portals have limited operating hours, which means that when customers require assistance outside of those times, they are often unable to speak to a support agent.

The emergence of Agentic AI

Agentic AI is being explored to tackle complex cases of process automation and develop solutions that would allow for semi-autonomous, automated work. The applications of agentic AI are extensive – many industries can benefit from automation work – and it’s often a matter of understanding what’s necessary to make it work.

Take, for example, a home loan application – an area of financing that often involves reading and interpreting large amounts of structured and unstructured documents to understand an individual’s borrowing capacity. An AI agent, acting autonomously, may be able to augment existing work by extracting essential details from financial documents – without additional human support.

For customer service tasks, agentic AI also holds substantial promise as a tool that can help supplement teams’ work. Agentic AI could help validate and verify personal information and even perform queries such as account updates, statement generation, and payments, allowing human support agents to be redirected to work on complex customer issues.

The sky’s the limit for agentic AI. As it continues to evolve, there’s no doubt it’ll make its way into use cases that we can barely fathom. Consider a future where you can get a health insurance quote from an AI agent – or set one up to complete your weekly grocery shop across many different websites. Each of these use cases could change how we use the Internet – in ways that could make it more efficient and helpful.

The importance of governance

Agentic AI is not without its challenges, however. Organisations must be aware of the risks and limitations of any new and emerging technology. Take, for example, an AI agent completing a mortgage application – what procedures and safeguards are in place if an agent incorrectly rejects an application based on a poor interpretation of the data?

What processes, policies, and procedures must be in place to ensure that agentic AI is effectively governed and monitored? While governance can often feel like an afterthought when human outcomes are involved, it’s crucial that it is enforced and there to protect both the organisation and the end users of agentic AI products.

As can be seen with previous releases of AI products, poor governance can often lead to poor outcomes. A classic example occurred not long after the release of ChatGPT when employees in the monitor arm of major tech manufacturer Samsung found themselves in hot water after posting sensitive corporate secrets on an open AI product.

A new age of online experiences

We’re rapidly entering a new age on the Internet – where our online experiences may be supplemented or even replaced by AI agents. Perhaps, one day, the only interactions we’ll have with real people online will be through emails and social media.

For those currently studying for qualifications such as an MBA, understanding how their knowledge and skills could be used to bridge the gaps between modern Robotic Process Automation (RPA) and agentic AI could not only meaningfully improve these applications – it could change the world.

Agentic AI opens up a world of new opportunities – in ways seldom imagined when Marconi’s radio hit the airwaves more than a century ago. These exciting innovations make you wonder – what does the future of AI look like in today’s constantly evolving world?

By Josh Young