Artificial intelligence (AI) can lead to opportunities for many businesses. But simply adopting the technology isn’t a magic solution to achieve growth and efficient operations—it takes significant planning, investment and strategic focus.
BMO recently hosted an event with three panelists who are experts in the field:
Gene Moo Lee, Associate Professor of Information Systems & Analytics, University of British Columbia Sauder School of Business
Marc Low, Director of Innovation, Growth and Emerging Technology, KPMG Canada
Jordan Atkins, Vice President, WTC Group, a leading logistics and transloading service provider
They offered their insights and experiences about how companies can effectively incorporate AI applications and tools through a three-step approach:
Awareness. Understanding specifically which parts of your business you can improve with AI.
Strategy. Determining how you’ll use AI to enhance your business goals.
Implementation. Understanding the details about how AI will be incorporated into your workflows, ensuring data integrity and security, and creating a culture of experimentation.
A summary of the event follows.
Developing awareness of AI’s potential
Before WTC even considered AI tools, the company took an initial but important step of making sure it had the right data in place and addressed relatively small data-related challenges before zooming in on even more granular issues.
“We spent the better part of the first five years of our digital transformation allocating most of our resources to understanding what data we could collect, and then trying to structure that in such a way that it made sense for those practical applications we were trying to achieve,” Atkins said. “We started with trying to improve container handles per shipment, trying to reduce truck trips to the port, those sorts of things. And then, as we got into it and realized how much more potential there was, those first small wins informed the next few steps in our journey.”
Building a business-aligned strategy
Once a company decides to implement an AI tool, the next step is to build a strategy that aligns with your business goals.
“We saw it from the beginning as a cost-saving or an efficiency play,” Atkins said. “Creating a more sustainable supply chain for us and our customers. We went for the low-hanging fruit to start—automating simple tasks, collecting, aggregating and leveraging the data to improve our decision-making at the operations level. And then we worked our way up the chain from there.”
When making those types of strategic decisions, Low said those discussions have largely been driven by IT departments. The most effective strategies, however, develop when company leadership is involved from the start.
“In a perfect world, it's a board and C-suite conversation to start,” Low said. “We're starting to see those leadership groups coming to the table saying: ‘This is transformative. Now what do we do about it?’ IT has to be part of that conversation to enable it, but it has to be driven at a leadership group level to try to understand what you're trying to accomplish, and then set the course for the right tools to get it done.”
Strategic planning also involves understanding all of the potential outcomes of implementing AI. “In knowledge work, if you unlock 15% to 20% of utility inside your organizations like KPMG did when we launched our ChatGPT tools, the harder question for leadership is what do you do with that 15% to 20%?” Low said. “Is my business going to be twice as big revenue-wise, or half as big headcount-wise? That's the fundamental question for this era.”
Effective implementation
While there are several commercial AI models that companies can use as the foundation for their implementation, WTC opted for an open-source model so it could manage the security and privacy of its data internally. Regardless of which model you use, Atkins said ensuring quality data is the first step to an effective implementation.
“If you don't have clean, good data, it’s garbage in, garbage out—and AI doesn't fix that,” he said. “That data piece is of core importance, and it takes the most time to get right. We spent a lot of years playing tug of war between the engineering team, the data analytics team, and the R&D team talking about how we were going to structure this, what was going to be the most valuable data. Having that solid base upon which to apply the AI to then leverage into your KPIs or whatever goals you've set as an organization is the very first and most important step that you have to get right from the beginning.”
As for what constitutes good data, Lee said it starts with asking a simple question. “What am I going after?” he said. “And then you start collecting the relevant data.”
Just make sure the questions you ask are the right ones. As Low said, “Don’t optimize a process that shouldn’t exist in the first place. If you’re asking the wrong question, you’re just moving faster in the wrong direction.”
You’ll also need to consider some of the potential unintended consequences of AI implementation. Lee pointed to cybersecurity and confidentiality concerns.
“Data is your asset,” Lee said. “One way to secure your asset is to have your own internal AI tools. Many companies are using open models on their systems and fine-tuning the model for their internal purposes.”
Another key to implementation is removing the barriers to adoption. As Low noted, resistance to change from within an organization is often a major roadblock when deploying any new technology. But he pointed out that there's something unique to generative AI that builds resistance.
“Generative AI is now in everybody's hands, and it doesn't take a rocket scientist to figure out it's going change the work that we do,” Low said. “And the work that we do pays our mortgages, pays our school fees for our children. So, part of the story is how do you build up a safe environment for your teams to experiment with it?”
For Low, it comes down to creating a culture of curiosity and experimentation. “As leaders in the business, you have to be proficient in those tools and driving that change,” Low said. “That begets curiosity inside the organization. And that's what rallies people around it. Make it safe for people to experiment and fail. Make it safe for people to bring ideas to the table.”
While AI promises great transformation, in many ways, it’s no different than adopting any other new business process: you need a clear strategy that defines what you intend to accomplish, and you have to bring your employees along on that journey.