As artificial intelligence threatens to upend the consulting industry, the head of Boston Consulting Group’s Canadian arm isn’t afraid of letting her top performers feel a bit of peer pressure.

Kathleen Polsinello leads BCG Canada from the 48th floor of a Bay Street skyscraper in downtown Toronto and helps clients implement AI systems. She knows BCG, and her competitors across the street, are being pushed to deliver clients real results, not just slide decks or reports.

“There’s many people that can deploy AI: our friends downstairs at Microsoft and across the street at many of the firms,” Polsinello said in an interview, adding that BCG is often pressed on why consultants are necessary for projects like building custom AI bots. 

Like many business leaders in white-collar industries, Polsinello must balance managing workers’ adoption of AI tools with ensuring those tools don’t eventually replace them. Management analysts are among the 40 careers that Microsoft researchers expect will be most impacted by AI. 

Still, Polsinello said she worries the firm’s senior leaders might be hesitant to learn AI-related skills from young associates. To address that concern, the firm now requires one Canada-based partner to open each monthly leadership meeting by presenting an AI-related win. Polsinello, who tracks the number of licences of AI products used by the team, has found that these presentations have become influential in encouraging other partners to try out new tools.

“That’s a bit of peer pressure,” said Polsinello. But, she added, the meetings also serve as “a bit of a safety net, of asking questions in a peer group.”

Boston Consulting Group famously has little patience for time-wasting projects, and was historically hired to tell clients which of their projects are “stars,” with high growth and market share, and which “pet” projects are “dogs” with both low growth and low market share. The firm is one of the three MBB firms—McKinsey, Bain and BCG—a group that is considered a small but elite alternative to the classic Big Four accounting and consulting firms.

Companies are taking different approaches to AI training. Rival PwC hosts “prompting parties” and challenges employees to compete in internal “game show-style competitions” to prove their AI skills. BCG, meanwhile, is eschewing internal hackathons and “play” projects, according to Polsinello. Instead, AI skills are diffused through an “apprenticeship” model where workers train each other on new tools as they move from project to project—even if that means a senior worker is learning from a junior associate. 

BCG faces growing competition, as rivals like Accenture chase AI talent and strike partnerships with hot startups like Anthropic, OpenAI and Cohere. With fewer than 600 workers in Canada, BCG must compete in areas where its workforce is far outmatched by Big Four firms like Deloitte and KPMG, which each have well over 10,000 Canadian workers. 

The Boston-based group has also faced scrutiny over two former U.S. partners’ undisclosed work on an Israeli-backed aid initiative in Gaza, which the firm disavowed. Meanwhile, North American governments slashed consulting budgets, and rival firms like McKinsey were pressed to drop work for foreign governments. BCG’s senior leadership in Canada reviews all its public sector work, Polsinello said. 

Polsinello said BCG Canada is taking steps to stay ahead of the competitive demands. The firm often gets paid variable fees based on the clients’ returns on a project, and Polsinello said BCG’s experience running pilots and providing measurable results to clients has helped it justify its work in AI. 

“Over the last 18 months, we have put a massive push on ensuring habitual [AI] usage,” she said. “I get a report every two weeks of all the habitual usage in this building and the one in Montreal and Calgary by cohort.” 

Still, firms like Deloitte have endured widespread criticism for misusing AI and giving clients reports containing AI hallucinations. Polsinello said BCG Canada has safeguards in place to ensure that AI is used for operational tasks—like keying annual reports into Excel or creating slide decks—rather than for creating narratives or analyzing compensation tables.

Seventy per cent of the value of deploying AI comes from “the change management of humans actually using it,” said Polsinello.