Last week, Accenture named Jodie Wallis its managing director of AI in Canada. In the new role, Wallis is responsible for overseeing Accenture’s investments in the Canadian AI ecosystem, growing its presence in the space, attracting AI talent and delivering that expertise to clients. Wallis is an Accenture veteran, having been there for 24 years, most recently as managing director of its Canadian banking business.

Earlier this week, strategy spoke with Wallis about just how much AI can contribute to Canadian business, what kind of talent the market still needs and things that need to be prioritized to take full advantage of the opportunity it presents.

How big is the opportunity in AI, specifically when it comes to business growth in Canada?

With every new innovation, we seem to say it’s going to be the biggest opportunity ever, but we’ve done research that backs up the fact that AI has that potential. We modelled two scenarios to figure out the growth rate for 2035 – one just using the current set of technologies and one with the same industries using AI. There was a massive difference. There’s a huge opportunity here for our clients, but also a complex challenge to help them figure out how to invest in AI to maximize returns.

What are the next steps you want to take in order to really capitalize on these opportunities?

One thing we’re starting to do is help our clients think about how to get value on AI and really see the business potential. We use the term “return on algorithm.” If you invest wisely in AI, which has a self-learning component, these are assets that appreciate over time compared to everything else we invest in, which depreciates over time.

I just had a conversation with a call centre client, and we talked through a scenario where AI could not replace, but could augment, the ability of the human agents to respond to a call quickly and effectively. So an AI system could listen to the question someone is asking, and before the agent even has to type anything, natural speech recognition could figure out exactly what the intent of the customer is and provide the answer on the screen to the agent.

Another example is in wealth management. There’s a certain amount of review of communications between an advisor and customer, and they look for things like overly pressurized sales tactics or insider training. This has usually been done with keywords, but we’ve developed an application that can integrate different data sources that obtain better insight about what bad behaviour really looks like and reduces the number of false positives that come out of the email surveillance. That increases the level of trust at the organization and, by extension, the customers.

With all the investment there’s been in the Canadian AI ecosystem, are there any areas of improvement you think the marketplace could be taking better advantage of?

The thing we have going for us in Canada is a really strong research capability, both in terms of the calibre of researchers we have and the investment being made in them. [There is] momentum in the space right now; we get calls from our international offices all the time about all the financing in the space.

But the first improvement we could make is getting the talent to extend outside of academia. We have these great researchers, but researcher talent doesn’t scale. There are only ever going to be so many world-class, pioneering researchers. So the question is, what do we do to get a set of skills leveraged from that base of academic excellence into people that are working in enterprises of all sizes, and get people who have expertise that might not be in the development of AI but can be useful in its application?

The other thing is money in AI.  We need all that money in research, but the adjacent capability we need to develop is the ability to commercialize it. That’s how we make sure it’s not just a few people who are researching and deep in the knowledge that can use it. If we commercialize it, it’s available to any company that has the resources to take advantage of it.

Is there anything besides talent and money that the market needs?

A lot of countries are advancing their data policies around AI, and Canada is behind a lot of them, certainly when it comes to the EU. Most AI algorithms need a really high amount of data to be effective, so the question is how can we pool our resources as Canadian companies to create the data sets that will further power AI? Otherwise, we’re sub-optimizing, and all I have is what is available within my company and what’s publicly available. We don’t really have a Canadian solution for building super data sets.  The other thing AI needs is computational power, so we need to advance our policy on moving this power forward in a way that can scale.

Are there any other ways Canada is behind other markets?

We are a little bit behind in terms of developing a framework around governance, and that’s not just cracking down when it comes to data privacy. We need real policies that will stimulate the development of AI and have the right controls in place. If left unchecked, AI can develop bias in its algorithms. We need to get ahead of that as Canadians so we are sure the amounts we’re spending aren’t going to lead to outcomes we didn’t ask for.

This interview has been edited for length and clarity.