EY presented a platform for finance and accounting work at the ninth China International Import Expo (CIIE). The Chinese newspaper Wenhui Bao reported the launch on 8 November on its Shanghai Observer service, and Sina Finance republished the story.

Three agent layers and shared orchestration

An AI agent is a program that uses a model to carry out the steps of a task. The platform has 15 agents, one orchestration hub and four core capabilities, the newspaper reported. Orchestration generally means that a system assigns tasks to agents and links their work steps together.

The agents are organised into three layers. In accounting work, they assist with recurring tasks such as expense review, bank reconciliation, voucher preparation, management of receivables and payables, and annual report checks.

In the analysis layer, the agents produce analysis reports and help users retrieve data, identify anomalies and understand changes in the business. In the business outlook layer, they support revenue forecasting, budgeting, cash flow analysis and simulation of business decisions.

The platform’s foundations include a shared data and knowledge base and a semantic layer. A semantic layer generally defines what financial concepts mean in the data of different systems.

Zhao Jianyun (赵剑澐), EY Greater China leader for technology risk and data intelligence services, said that ‘the value of finance and accounting AI lies not only in improving the efficiency of individual tasks, but above all in organically combining data, professional expertise and business scenarios’. Zhao believes AI must be built into real workflows rather than left as standalone tools, the newspaper reported.

The description is based on EY’s own presentation. The story includes no measurements of the platform’s accuracy or of working time saved.

Report: autonomy to match risk

At the expo on 8 November, EY also published a report on the use and trustworthy governance of AI in finance. A Chinese financial news service reported on it, and Securities Times (证券时报) and Sina Finance republished its story.

The report says finance AI progresses through four stages: rule-based automation, intelligent assistance, agent collaboration and decision support. AI autonomy should match the business risk and whether an action can be reversed. Organisations move to the next level by redesigning workflows, not by adding tools on top of old ones, the report says.

The report cites a global EY survey in which 68 per cent of chief financial officers said the way company value is measured needs urgent reassessment. Eighty per cent of them expect AI-driven business models to be present to a significant or moderate extent in their organisations within the next 12 months. The story does not give the number of respondents.

Neither story says when the platform will be available to clients or what it will cost.