Data & AI platform by Data Analytics

AI for accountants with room for professional judgement

Accountants have financial information that can prompt better questions for clients. The challenge is to organise that information carefully and assess findings in the right context. Depending on the application chosen, AI can support work and create opportunities to develop advisory services further. Professional judgement remains yours. Data Analytics helps configure the data foundation and its practical use.

The examples on this page are hypothetical situations, not customer cases.

Meet our team

At Data Analytics

Support the work, not replace accountability

Start with recurring tasks in which finding, comparing or preparing information plays a clear role. These might include exploring discrepancies or preparing discussion points. Determine what AI may support for each task and what review is required. A draft explanation differs from reviewed advice provided to a client.

Assess support for its usefulness within your workflow. Does the result agree with the underlying data? Are exceptions visible? Can a colleague follow the reasoning? Also clarify when an employee must not build on an answer. This prevents plausible wording from quietly taking the place of substantive assessment.

Separate portfolio views from client analysis

A portfolio view helps your firm identify where attention may be needed. A client analysis supports a conversation about one business. These purposes require different levels of detail and access arrangements. Employees using planning overviews do not automatically need access to all underlying client information. Configure these perspectives deliberately and separately.

Example: a portfolio overview shows which files lack current figures. That is a different signal from a substantive conclusion about those clients' financial position. Comparisons between businesses also require appropriate definitions, periods and context. A difference between two companies is not necessarily a shortcoming in either one.

Record data agreements for each use

Discuss why client data will be used, which sources will be connected and who will have access. Use for an individual advisory conversation may require different agreements from analysis across multiple clients. Have relevant contractual and legal questions assessed by appropriate specialists. Do not assume one general permission covers every future AI use.

Include data freshness, corrections and termination of access. Who corrects an incorrect connection, and how is a changed agreement implemented? When an external model is used, data processing depends on the model and configuration. Investigate the actual data flow and applicable agreements before making client information available in an application.

Use insights as the start of advice

Example — illustrative advisory conversation: you see that revenue is increasing while gross margin is declining. You ask the business owner whether the product mix, purchasing prices or discounts have changed. The owner then explains a temporary introductory promotion. The numerical signal helps open the conversation, but the business context determines its interpretation.

You can then determine together which additional information is needed and when to review the situation again. Do not present a possible next step as though it follows automatically from AI. State assumptions, uncertainties and the client's role. This allows analysis to support your advisory relationship without replacing the judgement of the business owner and accountant.

Organise a repeatable way of working

Choose a limited initial application with a responsible file owner and regular review points. Record definitions and checks so colleagues use the same assumptions. Collect questions arising in practice and adapt the workflow accordingly. Assess not only the text produced, but also the quality of the conversation and the follow-up recorded.

ENABLE is the Data & AI platform from Data Analytics for connecting and structuring business data and providing controlled access for dashboards, analysis, apps and AI. Data Analytics supports setup and management. For financial dashboards, see the existing page; this page focuses on your firm's working methods and the development of advisory services.

Frequently asked questions

Can AI finalise client advice independently?

Treat AI outputs as support requiring appropriate review. You need to be able to check source data, context and assumptions. Those providing advice remain responsible for their own work and the review approach they choose.

May we compare data across clients?

That requires a separate assessment of purpose, agreements, access and applicable obligations. Also investigate whether the data is substantively comparable. Portfolio-level analysis does not automatically justify access to identifiable information about individual clients.

Does every file need the same sources?

Not necessarily. Align sources with the question and make differences visible. Comparable analyses do require shared definitions and sufficiently similar data. State when missing information limits a comparison or conclusion.

How should we assess an initial trial?

Use files checked in advance and specific tasks. Assess factual accuracy, traceability, required corrections and usefulness for the client conversation. Also record when employees reject an answer; this provides valuable information for further configuration.

Discuss your question with Data Analytics

Which sources, users and checks does your application need? Discuss your question in an ENABLE demo.