“Don’t do AI for the sake of AI,” says Danielle Gifford, managing director of AI at PwC. 

“Really make sure what you’re leveraging it for in terms of the application, and understand what are the metrics you’re looking at.”

Speaking with Digital Journal at Inventures 2025, Gifford offered a sharp diagnosis of where Canadian small and medium-sized businesses (SMBs) stand on artificial intelligence. Many have already moved from experimentation to deployment. The new challenge is understanding how to optimize these tools, measure their impact, and protect data along the way.

This matters now because AI adoption is no longer confined to large enterprises. SMBs are seeing real returns on productivity and efficiency, even as they wrestle with privacy, integration, and workforce readiness. Gifford’s perspective adds an institutional lens to Canada’s evolving innovation landscape. AI has moved from buzzword to business priority.

[Watch the interview in full below]

Adoption is up, but strategic use is lagging

According to Gifford, roughly 80 percent of Canadian SMBs have begun exploring AI tools, and about 40 percent have integrated them into their business strategies. A Microsoft survey from March 2024 aligns with this, finding that 78 per cent of Canadian SMBs are currently looking to adopt AI, and 65 per cent are encouraging their employees to use it. 

“They’re seeing a pretty significant outturn in terms of ROI,” she says. “I believe the stat is about 127 percent in terms of productivity gains and cost takeout in terms of what the actual applications are.”

But those wins are not evenly distributed. 

Many businesses still approach AI in a fragmented way, without clear guidelines or measurable objectives. The most common use cases involve back-office functions like finance, operations, HR, and marketing, she explains. These are areas where repetitive tasks make automation more impactful. These gains often come before deeper questions around governance, ethics, or long-term scalability are addressed.

As more tools embed AI directly into their platforms, such as CRMs or payroll systems, Gifford cautions leaders to look closely at what is already available before buying new solutions.

“Knowing and understanding your tech stack and just enabling and turning some of those things on when you can, makes much more sense than trying to build or buy a tool off the shelf,” she says.

Optimisation means new thinking and cross-functional teams

Once AI is deployed, the next hurdle is optimization. That is where strategy, governance, and employee experience converge. For Gifford, too many businesses underestimate how people and policies need to evolve alongside the technology.

“If you’re just giving your employees ad hoc access to ChatGPT or to Grok or to Claude or to Loveable, are they giving it sensitive information, like personal data or client data?” she asks. “Are they giving it your company financials?”

In regulated sectors like finance, energy, or healthcare, Gifford has seen clients stipulate in contracts that data used by vendors must reside in Canada or even in specific provinces. The question of where data lives, how it is stored, and whether it trains external models is no longer abstract, it is a growing point of negotiation.

Leadership is also being reshaped: “Whether you like it or not…seven out of 10 employees are using AI, and so if you have a stance as a company — the do’s and the don’ts of when to use it, where to actually apply it — I think it’s really helpful for your employees to know and understand.”

To that end, Gifford has also noticed that some companies are setting up AI councils or task forces, bringing together HR, legal, risk, finance, and business units to evaluate tools and set internal policies.

“It’s not meant to be some rigid governance,” Gifford says. “But it’s really meant to be a committee that has representation [from] HR and finance and legal and risk, and some of your business unit leaders so that they know and understand some of the different use cases.”

One point she emphasises is the inclusion of HR in these discussions. Too often, people teams are left out of AI conversations, even though employees are among the most affected. Fear of job loss, confusion about how to engage with new tools, and questions about upskilling are real and persistent.

“Your people are some of your biggest assets,” she says. “There’s so much noise out there. People are really scared. Is this gonna take my job? Is this something I’m gonna have to compete with?”

Keeping the momentum 

As the hype around AI stabilises, Gifford is focused on how companies maintain focus and extract value. That means applying the same scrutiny to AI subscriptions and tools as they would to any business investment.

“If you’re paying for [Microsoft] Copilot licenses, and it’s $40, $50, $60 a seat, it all adds up,” she says. “So really making sure that when you’re applying AI, you understand the application, and you have those benchmarks in place to know that you are getting good value out of it”

Beyond metrics, keeping AI visible across the organisation is also critical. Gifford recommends putting AI on agendas, celebrating wins, and treating deployment not as a project but as an ongoing evolution.

That evolution includes revisiting systems, tools, and capabilities on a regular basis, and above all, not letting it fade into the distance.

“Keeping it top of mind, keeping it on agendas, making sure that people know and understand that they’re heard.”

A leadership lens on responsible adoption

Gifford’s message is grounded in pragmatism. Canadian SMBs are not in a holding pattern. They are acting, adapting, and iterating. But as AI moves deeper into operations, leadership must sharpen its focus on long-term strategy, cross-functional governance, and clear ROI.

Her advice reflects a broader shift in Canada’s innovation landscape. AI is no longer a separate thread. It is becoming part of the national business fabric. Whether through policy, procurement, or partnerships, the next phase of innovation will be shaped not just by what tools are used, but by how and why they are adopted.

“Tech stacks are changing, they’re evolving,” she says. “Making sure that your tools are really working with you and not against you is going to be something that’s really important.”

“Make sure your tools are working for you,” Gifford says, “not just creating more noise.”

Watch the interview: