Dominic Lam and Stephen Perelgut · uOttawa PDI News ·
Canadian businesses are notoriously conservative and risk averse. They often try to stick with what works and avoid unfounded new ideas until they’ve proven their merit. There’s nothing wrong with that, and often it’s the right approach. But it’s the wrong approach when it comes to AI. It’s here, and if your business isn’t figuring out the best way to make use of it, you’re already falling behind.
However, that’s not to say that you should throw caution to the wind. Using AI for business is not without risk and the technology has not developed to the point that businesses can simply hand over vital processes. Integrating AI into operations must be approached strategically. Nor can AI be treated as a set-it-and-forget-it solution. The technology is constantly changing, and quickly, new iterations of AI with new capabilities are emerging, and it’s crucial to keep pace.
To compete in this constantly shifting landscape, businesses must understand not just what AI can and can’t do well, but how that fits into their specific business needs, and the risks they face by adopting or not adopting.
AI Is Shifting From Reactivity to Proactivity
Much of the AI currently available and being used by both consumers and businesses is what we call augmentative intelligence, rather than true artificial intelligence. It’s meant to work in concert with human beings to perform specific functions. ChatGPT for example is used by many people as a query-based search engine. Essentially, users ask questions and work with the answers.
The current trend is moving towards AI agents. Rather than being reactive interpreters of inputs, more sophisticated algorithms are emerging that can take a broad set of instructions and execute on them proactively. This is closer to a true AI that is capable of conducting its own activities without human input, beyond the initial request.
The significance of this from a business perspective is that AI is rapidly shifting from being an option that can be used to replace personnel that perform a specific function, to one that can actively undertake new business processes. While AI is currently being adopted to help businesses reduce costs, the emergence of AI agents will allow businesses to use technology to generate additional revenues, by expanding existing revenue streams and seeking out new ones. This competitive advantage is one of the fundamental reasons for urgency. Businesses that don’t successfully adopt AI agents will simply begin falling behind.
Benefits and Risk Are in Equal Proportion
One of the major business benefits of adopting existing AI systems is cost savings. After paying an upfront cost, many AI technologies tend to run more cost-efficiently compared to having people perform the same tasks. They don’t collect wages, take sick days or lunches, or even sleep.
New emerging models take this one step further. While Large Language Models (LLMs) like ChatGPT are the best-known, emerging Small Language Models (SLMs) may be better suited to businesses. These models build on smaller, more specialized materials directly related to the industry they’re meant to operate in. This makes them both more cost-effective and efficient than the more general purpose LLMs.
But while more industry- and business-specific AI systems offer more opportunity for cost savings and the promise of potential revenue generation, the greater the integration, the greater the risk. This is the cost-benefit analysis that every business must do before adopting existing or emerging models.
AI Must Be Used But Cannot Be Trusted
All AI systems are very good at creating the impression that they think the same way we do. But the reality is that they don’t. Nobody really understands the exact way they arrive at conclusions or make some decisions, even if the processes are understood in general. Academic research in this area is in its infancy.
Still some businesses are simply asking an AI program what to do, or what the answer is, and taking the answer completely at face value. This is dangerous. Air Canada learned that the hard way last year, after its customer service AI gave a customer wrong information. The settlement value was low but illustrates the point that simply handing over critical business services to an AI is not without risk. While AI systems are becoming increasingly sophisticated, they are very far from being infallible. And larger scale integrations magnify this risk. An organization like Air Canada deals with tens of thousands of calls a day. If an AI is hallucinating an answer 1,000 times a day, that can quickly become a very costly problem.
But integrations of AI into critical processes continue. One of the largest adoptions, and probably one of the biggest integration gambles is a recent move by Microsoft to bring AI in to write even more code and also handle Quality Assurance (QA) in the coding pipeline. This is an ambitious step. Time will tell if the AI is truly capable of both producing code and ensuring its quality, especially on the scale needed by a company like Microsoft.
Understand Both AI’s Capabilities and Your Business Needs
Successfully adopting AI for your business is all about looking at both what AI can do for your business on its own and how it can augment the role of the people in your business.
This isn’t easy, but there are some best practices. One of the most important things to consider is whether you’re asking the best questions, or even the right questions. Prompt engineering is a rapidly growing field designed to tackle this exact issue. Similarly, leveraging industry-specific insights and knowledge, both to guide better prompts and to further develop what AI gives you, is a crucial differentiator. Even if everyone in your industry is using the same SLM, the knowledge and experience of the people in your business will provide a competitive edge that can’t be replicated.
To do this successfully, it’s necessary to bring the two worlds involved — technology and business — together. The technical folks who deeply understand how to deploy and make use of AI tools, and the business managers with the marching orders about what AI needs to accomplish, must be able to communicate with each other successfully.
Turn AI Into a Competitive Advantage With uOttawa PDI
This is why we created the Practical AI for Business Management program at the University of Ottawa’s Professional Development Institute (uOttawa PDI).
This program walks business managers through the fundamentals of AI, from its basic mechanisms and its potential value through the pitfalls and dangers it poses. Real-world case studies from multiple industries highlight how AI can and has been used to enhance business strategies and improve decision-making. Participants learn how to identify the most critical elements of successfully adopting AI, including effective sponsorship and stakeholder engagement, culminating in a hands-on case study that guides participants through the process of developing actionable AI strategies targeted at specific business needs.