From bookkeepers to business advisors: How AI Agents are reshaping accounting

The conversation around artificial intelligence in finance has moved decisively from aspiration to implementation. At The Future-Ready Firm: Scaling Your Accounting Practice with Cloud Solutions, a webinar hosted by Intuit QuickBooks and CorporateTreasurer, three practitioners laid out what that transition looks like on the ground and why the window to adapt is narrowing.

The case for agentic AI 

Eric Chan, founder and chief coach of True Financial Training Academy and a QuickBooks Pro Advisor, opened the session by drawing a sharp contrast between traditional rules-based automation and what he called "agentic AI", a system that doesn't simply execute preset instructions but learns, infers and acts with context.

"It is not about replacing accountants," Chan said. "It is about letting them focus on what actually matters."

The distinction, in his framing, is consequential. Where conventional bookkeeping software waits for a user to notice a new vendor and set up a categorisation rule, QuickBooks' suite of AI agents, covering accounting, customer management, finance and project management, identifies the pattern autonomously. If a business pays AWS for the first time, the accounting agent suggests "cloud services" as the category, learns from the approval and applies that logic to every subsequent AWS charge without prompting.

Anomaly detection works similarly. Rather than requiring a staff member to scan hundreds of line items looking for outliers, the finance agent surfaces a flagged transaction proactively. For example, an office supplies invoice at ten times the usual amount, with context attached. The system does not act unilaterally; all suggestions pass through human review before being finalised.

"Agent tech AI is not a future concept," Chan said. "It is already inside QuickBooks today."

For Chan's own firm in Hong Kong, the shift has been measurable. Staffs spend less time on transactional data entry and more on advisory work, with revenue mix shifting accordingly toward consulting. This he says is a higher-margin offering than traditional bookkeeping or audit.

The augmented accountant 

Lily Tan, FCCA co-founder and compliance director of TnB Global Outsource and a QuickBooks Pro Advisor based in Singapore, used her session to introduce what she called the "augmented accountant", a professional whose value lies in judgment and interpretation rather than data handling.

"AI will not replace us," Tan said, "but it will replace persons who do not adopt the technology."

Her firm manages approximately eight companies across different industries, all running on QuickBooks Online. A real-time anomaly picked up by the finance agent recently flagged a material inter-company billing discrepancy. It was a case of revenue booked in one entity without a corresponding cost entry in another, that Tan acknowledges was missed despite her experience. The catch directly affected profitability visibility.

The framing she offered: adaptability, preparedness, collaboration, or ABC, cuts to the core of the transition challenge.

Her four-phase model maps the journey from digital adoption to full advisory practice: digitisation first, then automation, then predictive forecasting, then strategic advisory.

Many firms in the region, she noted, remain anchored at phase one or two, still relying on Excel-based processes that have not fundamentally changed in decades.

"If we don't change now," Tan said, "the world is going to change without you."

Cloud as an operating model 

Michael Aguirre, founding and senior partner of UHY M.L. Aguirre & Co in the Philippines and a QuickBooks Pro Advisor, grounded the discussion in a client case study: a boutique 20-room hotel in Boracay operating from a head office in Manila.

Before moving to QuickBooks Online, the hotel ran its daily reconciliation process through spreadsheets emailed between locations - sales tabulated on site, submitted manually, matched against deposits by a finance team hundreds of kilometres away. Errors accumulated. Timing gaps created control risks.

With a cloud setup, on-site staff record transactions in real time. Manila validates against bank feeds the same day. Disbursement authority remains centralised. Month-end closing, Aguirre said, now happens on schedule, something the owner, a former American corporate executive with a strict reporting culture, had demanded from the start.

"Cloud accounting is not just a software change," Aguirre said. "It's an operating model that gives management real-time visibility, stronger control and faster flows across multiple locations."

His advice on where firms should begin was straightforward: start with the outcome the owner actually needs. "I would say start from the boss, it's really the intent," he said.

Invoicing discipline, disbursement tracking and journal automation, in his view, are the three pressure points that unlock timely reporting and once those are functioning, AI agents can meaningfully amplify the results.

Across all three presentations, one message recurred: the tools are available now, the learning curve is manageable and the cost of inaction is rising. As Chan put it, "We can just ignore it and be left behind or we can make use of it to increase our capabilities and serve our clients."