AI won’t fix your back office — it will inherit it.

Brian McGlynn is the CEO of Canadian investment portfolio accounting software company Wealth Write.Up.

AI ADOPTION across accounting firms quadrupled in a single year, rising from nine per cent to 41 per cent. Yet only 37 per cent of firms are investing in AI training for their teams (Karbon, 2025). That gap is not a technology problem. It is a sequencing problem, and it matters most for the accounting professionals managing high-net-worth client portfolios.

The mainstream narrative is easy to absorb. Ninety-three per cent of accounting firms now offer advisory services, up from 83 per cent a year ago. Automated reporting adoption in family offices climbed from 46 per cent to 69 per cent in a single year. The direction of travel is clear, and the firms not moving are falling behind.

The back office tells a different story.

Many Canadian accounting firms, from national firms like MNP to regional practices and sole practitioners, now run family office and high-net-worth wealth advisory services. These are the practices where the gap between AI enthusiasm and operational readiness is most acute, because the data they depend on is the hardest to get right.

According to the Simple 2025 report, 57 per cent of family offices still rely on spreadsheets for critical investment tracking and reporting. More than a third manually aggregate financial data across disconnected systems. If your firm advises these clients, their data problem is your data problem.

Private market assets now represent 45 to 50 per cent of the average family office portfolio. That data does not arrive clean, connected or audit-ready. For the accountant preparing reports, reconciling statements and answering to regulators, this is a daily operational burden.

Nowhere is this more visible than in the investment subledger, where positions, transactions, income allocations and cost basis are tracked across a portfolio. In most family offices, that subledger lives in Excel or does not exist as a formal system at all. Automation success depends more on data structure than on AI capability.

Intelligent tools need something intelligent to work with. A model sitting on top of siloed spreadsheets, manually reconciled ledgers and inconsistent naming conventions is not producing insight. It is producing confident-sounding noise.

The governance gap makes this more urgent. A Zscaler 2025 analysis identified 4.2 million data loss violations across consumer AI tools, including platforms that staff reach for daily without formal approval.

Three-quarters of employees using unapproved AI tools have admitted to sharing sensitive information, most commonly customer data and internal documents. For an accounting firm managing the private financial lives of high-net-worth clients, that is not an IT incident. It is a potential breach of fiduciary duty.

IBM’s 2025 Cost of Data Breach Report puts the cost of an AI-related breach at more than US$650,000, with shadow AI incidents carrying an additional $670,000 premium. Caution among partners and principals is not backwardness. It reflects a real risk that the enthusiasm of the sales cycle tends to gloss over.

Accounting professionals serious about capturing the AI advantage need to complete the pre-work before signing any vendor contract. Start with a clear-eyed audit of the back office. Where does data live? Who touches it? How is it reconciled? Could it survive a regulatory review?

Move critical financial functions off spreadsheets and onto purpose-built platforms, beginning with the investment subledger. It is the foundation every downstream report, tax calculation and performance analysis depends on.

Establish a written AI governance policy that defines which tools are approved, under what conditions and with what data. Invest in enterprise-grade systems with documented data handling agreements, role-based access controls and clear commitments on whether client data is used for model training. Consumer tools built for general use have no place in an environment handling private client wealth.

The AI opportunity for accounting firms serving high-net-worth clients is genuine, and the competitive advantage for early movers is real. But early does not mean rushed. The firms that will capture it are the ones doing the unglamorous infrastructure work right now.

AI will not fix a data governance problem. It will inherit one.

Brian McGlynn is chief executive officer of Wealth Write.Up. Title image: iStock ID 679531800.