How CPAs in Small and Medium Enterprise Can Adapt and Thrive with AI
March 17, 2025
Artificial intelligence is reshaping CPAs’ work from the ground up, and small and medium practices (SMPs) are at the forefront of the transformation.
While large firms make significant investments in enterprise AI applications and market propriety models, SMPs—with lower operating costs and more agility to scale operations—are well-positioned to accelerate productivity and scale through strategic AI adoption.
By leveraging a mix of early-stage generative AI (GenAI) tools including Microsoft Copilot or ChatGPT to automate routine tasks, CPAs can focus on strategic initiatives, service enhancement, and relationship-building to drive their practice’s productivity and growth.
At CPA Ontario’s 2025 Corporate Finance and Controllers Conference, two CPAs and AI experts shed light on how AI is changing the profession today and practical applications of AI for SMPs.
The Evolution of AI Workflows
Erin Kelly, CPA, CMA, and President and CEO of Advanced Symbolic Inc. (ASI) illustrated the shift from technical AI use to GenAI large language models (LLMs) and emphasized the benefits of LLMs in automating routine tasks.
“In the early days, interacting with AI was a technical endeavour often limited to niche applications and specific tasks,” said Erin while describing the development of Polly, ASI’s patent AI that produces unbiased social media audience and market research.
“Our proprietary approach allowed Polly to achieve over 90 per cent accuracy,” she continued. “The emergence of LLMs has enabled us to develop consumer-facing products.”
Tomas Milo, CPA, XBRL expert and PhD, Accounting candidate at McGill University, echoed Erin in calling ChatGPT’s November 2022 arrival in the market “an inflection point in how people can interact with and train AI at scale.”
Citing his PhD work in training machine learning models to detect irregularities in financial statements, Tomas agreed that the advent of LLMs significantly evolved the technical nature of AI to make it more accessible and consumer friendly.
But he cautioned CPAs against trusting or interpreting the output at face value: “Unlike Excel or Power Automate’s predictable responses, AI’s variance often misunderstands prompts or hallucinates.”
Levelling the Playing Field for CPAs
CPAs in SMPs have strategically integrated LLMs into their workflows, enhancing their teams’ efficiency with tasks such as document reconciliation and VLOOKUPs.
“Smaller organizations have the most to gain from AI, but using the right tool for the task is essential,” said Erin.
“GenAI is more suited to creative tasks and has significant limitations, while discriminative AI, which classifies data by distinguishing different categories, helps firms reduce costs while maintaining high levels of accuracy and efficiency.”
“The emergence of agentic AI systems capable of autonomous decision-making will revolutionize how CPAs work, freeing them up to focus on strategic and analytical tasks while AI manages routine processes."
Tomas predicted that SMPs who strategically invest in people and tools and consider both financial and non-financial aspects of adoption will thrive in a changing industry—particularly as open-source models make AI more accessible and cost-effective.
“Developing policies for the appropriate use of public tools to ensure data security and privacy, and training employees to effectively harness their full potential to ensure efficient workflows, are the most practical applications of AI for CPAs.”