From Generative to Agentic AI: Daniel Mortimer on the Profession’s Next Frontier
September 4, 2025
Daniel Mortimer will be speaking at CPA Ontario’s upcoming AI Conference Adapt, Evolve, Scale on September 18. Join us to gain actionable insights and real-world examples of CPA-centered AI workflows to bring your practice to the forefront of digital transformation. You can register here.
Artificial intelligence is a fundamental disruptor, transforming how CPAs work and the value they bring to business and the economy. But even as the profession adapts, the pace of innovation is not slowing down.
In just a few short years the technology has evolved from generative AI, which replies to commands, to agentic, which can set goals, plan and execute tasks. Agentic AI presents endless possibilities for elevating the work of CPAs and introduces questions about the ethics of its use.
Daniel Mortimer, CPA, CA, a Senior Manager at EY, is a leader in technological transformation and AI adoption. On September 18, he will join CPA Ontario’s AI Conference for a fireside chat on AI 2.0 and autonomous systems.
We sat down with Daniel to discuss agentic AI and how CPAs can thrive in an age of rapid automation.
The Leap: From Generative to Agentic AI
Generative AI has established its capacity to draft reports, summarize data and even write code. But the next wave goes one step further. It can make decisions, automate workflows and integrate across multiple systems, turning AI from a tool into a collaborative digital agent.
Daniel illustrates the shift: “Traditional AI could give you an answer like, ‘Here are all available flights on Air Canada.’ Agentic AI takes the next step - it books the flight. In accounting, that means more than citing a standard. It analyzes how it applies to a client, flags discrepancies and even suggests next steps.
For example, someone working in audit could input IFRS 10 into the platform and it would tell them how it applies to their exact situation and provide an in-depth analysis.
Daniel suggests that the next step will be firms moving away from AI that exists in third-party tech spaces: “I think creating their own interfaces will make it a little bit easier for our people to use. These platforms would inherently know you are an auditor, and maybe even know the client you are dealing with.”
Crucially, this doesn’t displace professional expertise. “This isn’t about replacing human judgment,” Daniel emphasizes. “It’s about enhancing it. Agentic AI becomes a thought partner – augmenting your work rather than drafting everything for you.”
The First-Cut Advantage
Artificial intelligence thrives on creating the “first cut” of content, like drafting memos, summarizing research, or analyzing defined data sets. Tasks that used to eat up hours are streamlined to minutes.
“You iterate with it,” Daniel explains. “It’s a collaboration. AI provides the first cut and you, as the expert, guide it toward the final product.”
He likens this to Excel’s role in accounting – ubiquitous, indispensable and time-saving – but with the potential to free professionals for higher-order work: creativity, judgment and strategy.
Guardrails: Responsible and Ethical Use
The risks, however, are real. Incorrect AI “hallucinations” can compromise accuracy if left unchecked. Daniel underscores the importance of responsible adoption, from thoughtful prompt engineering to anchoring AI to verified sources.
Data privacy is another critical concern. “The data is king,” Daniel stresses. “We must make sure any AI tool we deploy aligns with our responsibility framework. It’s not just about capabilities – it’s about trust and confidentiality.”
For Daniel that means taking a responsibility-first approach, even if it delays speed to market. The priority must be maintaining the integrity of both information and professional judgment.
Balancing Automation with Human Judgment
As AI becomes more powerful, the profession faces a key challenge: knowing when to lean on automation and when to rely on human insight.
“AI is a tool in our toolbox,” Daniel says. “It shouldn’t be used just because it’s the latest thing. The goal is the best outcome. Sometimes AI is the answer; sometimes traditional methods are better.”
That balance keeps professionals at the center as strategic decision-makers, not passive overseers.
Trust, Literacy, and the Learning Curve
Canadian research shows relatively low public trust in AI, driven by privacy concerns and unfamiliarity. Only 34% of Canadians report that they are willing to trust information generated from AI. Daniel argues that education is the antidote.
“Don’t be hesitant because of fear,” he advises. “Approach AI deliberately, understand its capabilities and adopt a ‘trust but verify’ mindset. This is like learning Excel or Word - it’s a skill set that enhances your value.”
For CPAs specifically, AI is already streamlining tasks such as reconciliations, variance analysis and report drafting. Professionals who adopt it early will gain a competitive edge. “It’s not going away,” Daniel notes. “The more comfortable you become, the more capable and marketable you’ll be.”
A Thoughtful Future
For Daniel, the future of AI is not about disruption, it’s about elevation. By combining human expertise with thoughtfully integrated agentic AI, CPAs can reclaim time, sharpen insights and expand their impact.
“Our responsibility is twofold,” he says. “We need to safeguard the sensitive information entrusted to us, while making responsible use of AI tools that can help us do our jobs better. That balance is the foundation of a future where AI and human judgment coexist harmoniously.”
The message for CPAs is clear: AI is here to stay. By approaching AI with the right mindset, CPAs can adapt and evolve with agentic AI to deliver more strategic value to business and the economy.