AI Insights: Opportunities and Risks

Disponible en anglais seulement

We all work in a dynamic environment. What remains constant in business are the changes we have to navigate, particularly when it comes to adapting to rapid changes in technology. 

We recently held an event to discuss the changes impacting business decisions today and well into the future. I hosted a roundtable on the impact artificial intelligence will have on organizations. We heard perspectives from technology experts from the corporate and investor sides. Our panel included: 

• Saaya Nath, Partner, Jump Capital, a technology-focused venture capital firm 

• Jimmy Paul, Chief Technology Officer, Co-founder, Crafty, a centralized platform for workplaces to manage food, beverages and supplies for their in-office, remote, and hybrid teams across the globe 

• Lawrence Wan, Chief Architect and Innovation Officer, BMO 

Following is a summary of our conversation.  

The AI explosion 

It’s almost mind-boggling to think about how quickly AI has come to dominate conversations about the future of business. Until last year, discussions around AI were mostly limited to niche areas of the tech community. Now, generative AI tools built on large language models, such as ChatGPT, Microsoft Copilot and Google Gemini, are in the headlines daily. Still, several misconceptions about what AI is remain prevalent. 

“At its fundamental form, AI is the ability for a computer or software to mimic what we as humans do, specifically intelligence,” Nath said. “AI is such a hot topic right now because it’s grown up. Models can now make much more complex decisions than they could 70 years ago, and it’s so pervasive now.”

AI includes tools as widespread and seamless to the user experience as recommendations on Amazon or Netflix, or digital assistants such as Siri and Alexa. The generative AI systems that we’re seeing today, Wan explained, relates to how creative the output can be.

“Before ChatGPT, a lot of conversations about AI were around solving very specific tasks,” Wan said. “The output was very narrow; the prediction was very narrow. But now you’re getting to a space where AI is trying to create something new, you can generate things instead of repeating things with a narrow output.” 

From an investor perspective, Nath said she focuses on what will enable the adoption of AI within organizations. Technology decision-makers within companies know AI will be an integral part of their operations. Overcoming the challenges to implementing AI within their organizations, Nath explained, comes down to understanding the risks and how to mitigate them, and determining what resources are required. 

“Computing power has come a long way, but it still isn’t necessarily where it needs to be for most organizations to adopt it easily,” Nath said. “It’s very expensive, it requires very expensive talent, a lot of infrastructure, and a vast amount of data that’s clean, mature and well beyond what most organizations have today. We’re excited about solutions that are helping to solve for those [concerns], because we think those will be the big movers that will enable mass adoption within organizations.”

Business use cases

All of the panelists pointed out that most of the use cases involve experimentation for improving internal productivity, such as sales and marketing teams using ChatGPT to generate note summaries or email drafts. Crafty, for example, uses an AI sales intelligence system called Gong, which records calls, creates transcripts and examines those transcripts to recognize specific themes, such as pricing. “We can start thinking about how we can change the way we price and make changes as an organization,” Paul said. 

For now, that’s the type of AI use case that will be most prevalent among organizations. Nath pointed out that there’s been significantly less AI adoption on the customer-facing side, largely because of the legal and regulatory risks involved.

Improving individual productivity on specific tasks is one thing. Translating that into bottom-line improvements from a revenue-generating perspective is a more difficult proposition.

“You’re trying to look for things that would fundamentally change your business processes,” Wan said. “What are the prerequisites we need to create for some of these things to work? Your system has to be in the cloud already, because that’s the only way you can create the on-demand capacity. Also, your business processes already have to be digital, otherwise it will be too difficult to create the necessary sets of data to drive these functions. As you go through that process, that’s where you can identify the relevant use cases for you. As you’re changing your processes, that will be the opportunity to identify what kind of new capabilities are available to you.”

The second-mover advantage

In many areas, businesses like to take advantage of being the first to a new way of doing things. But is there a first-mover advantage when it comes to AI adoption? Or would it be better to wait and see what works and what doesn’t?

Nath acknowledged that because the technology is evolving so rapidly, there may not be much of an advantage to being a first mover for many companies. But she also believes that companies that don’t experiment with AI early could be putting themselves at a significant disadvantage.

“I don’t think there’s an end state where generative AI will be fully developed in five years and you can say, now I can go build toward this,” Nath said. “Organizations that don’t start learning about it now will not be able to make up for that time in the next three years. That’s where the internal use cases around productivity make a ton of sense. They’re low risk and very high reward.” 

And that doesn’t necessarily require making a large financial investment. Experimentation could include having employees experiment with ChatGPT to assist with tasks outside their main job function. Paul himself used it to help design a workshop agenda for his data team. 

Paul noted that while most of the technology Crafty makes gives it a first-mover advantage in its industry, AI poses a more nuanced challenge. For Paul, AI adoption must go beyond using AI for the sake of saying you use AI.  

“What we’re trying to do is solve customer problems, and AI is a tool we need to consider when solving those problems,” Paul said. “It’s important to know the technology, understand it, know what it’s good at, what it’s not good at, so that when there’s a problem to solve, it’s something you can use.”  

AI and the future of work 

AI’s implications for the labor market have been heavily debated. Fears of widespread job losses across industries of all stripes usually top the list. It also means younger generations who haven’t yet entered the workforce may need to rethink how to achieve their career goals.

“Whatever your passion is, it’s going to be disrupted in some way,” Paul said. “You can still go into that passion, but just know that the toolset might be different.” 

Nath noted that the U.S. economy has always reacted positively to previous technological revolutions. While it forces workers to adapt, it’s also consistently resulted in new revenue opportunities. For Nath, there’s another big transformation that AI will bring. 

“It really changes what you need to be in the workforce,” she said. “Humans are inherently biased to ask, where did you go to school? What is your degree? Once AI is trained properly and learns from its surroundings, there’s an argument that it will make better decisions on the skills that are needed for you to be in the workforce. You might not need a traditional college degree; you might not need to go to an Ivy League school to get the best job. It can allow people who maybe haven’t had those traditional avenues of education to learn things that they couldn’t before and be in roles that they couldn’t have before.”