At a glance
- AI is accelerating the automation of routine accounting tasks.
- The focus is shifting from bookkeeping to reporting and forecasting.
- Conversational AI assistants are becoming the new user interface.
- The accountant’s role could well shift to supervising AI and advising clients.
Every year, accounting software becomes one step more automated than before. Software now updates bank feeds daily, auto-matches transactions, and auto-populates forms. Accounting software’s creators are working to fully automate bookkeeping and reconciliation, though we don’t know when those capabilities will arrive.
The explosion of investment in artificial intelligence (AI) is not just increasing the pace of accounting software’s improvement, but extending it into new activities.
In particular, accounting software is now developing reporting and forecasting capabilities. This is the layer where finance turns transactions into decisions.
Some of the biggest questions for business owners are why things are happening, and what’s going to happen next. Why did cash dip this month? What changed in receivables? If we hire in May, what happens to runway? And what would need to be true for margin to recover next quarter?
Vendors are racing to bolt conversational reporting onto their platforms and adjacent tools. Some are putting AI assistants inside accounting packages. Others are pushing intelligence into reporting and compliance layers, or into ERP systems where finance data intersects with inventory, payroll and operations. The common promise is the same: fewer forms and filters, more narrative explanation, more scenario planning.
In that world, platforms that stop at reconciliation and compliance start to fade into the background. The ambitious vendors want to become the trusted reporting brain of the business. If they can make this work, the AI will explain results, and model scenarios with clients. The role of the accountant will no longer be to build the report, but to supervise the agent that builds the report.
Your accounting agent
Accountants have always struggled to sell CFO-style services, because there has simply been too much compliance work to get through first. If AI reliably strips time out of reconciliation and other routine tasks, the time and economic incentive shifts back toward advisory services. Accountants will then spend more time explaining what is happening, surfacing anomalies, and then running scenarios with clients.
AI assistants can simplify and speed up these reporting tasks. Instead of clicking through screens, menus and modules, users can ask the assistant to run a workflow, pull information, recommend an action or generate a report.
Xero, for example, is developing an “AI financial superagent” called JAX – short for Just Ask Xero. According to Eli Danziger, Xero senior vice-president of growth and core accounting, this system is being designed to:
- learn how a business operates;
- automate routine tasks (with oversight); and finally
- surface insights.
(Danzinger provided a written response to questions from Public Accountant.)
JAX makes a big promise – that it will act as a fully capable digital “team member”, able to reconcile, forecast and lodge on your behalf. Such capabilities still feel some way off.
How AI is changing reporting
The systems cannot yet to all that Xero would like JAX to do. But chatbots are already playing supportive roles in reporting. Many tools can already query a transactional database and explain the results in plain English.
The near-term shift is from form-based reporting – choosing a date range, period and transaction type – to simply asking, “show me this month’s cash position and what changed”.
Danziger says JAX is intended to answer questions such as: “What is the cash flow impact if I hire staff in May?” – and without the user having to build a model or load new data. He says JAX will draw on Xero data, connected apps and internet research via OpenAI.
This focus on reporting fits neatly with the longer-term push to automate bookkeeping. As reconciliation becomes more automated, accounting systems shift from recording the past to testing scenarios and producing forecasts.
As part of its mission to develop better insights from its core transactional database, Xero last year acquired business intelligence tool Syft. Other accounting software firms are moving in the same direction.
As accounting software shifts from processing yesterday’s transactions to forecasting tomorrow’s performance, chatbots are becoming the way we run scenarios, test assumptions and decide what to do next.
Now ERPs can talk too
The idea of using a chatbot to talk to your accounting software raises an obvious question: what accounting software will you use? What if your AI assistant is clever enough to operate a full enterprise resource planning (ERP) system for you, rather than just small business accounting software?
Oracle NetSuite has been rolling out Ask Oracle, a natural-language assistant designed to help users search, analyse and act across NetSuite using plain English. That may be an indication of where small business accounting software is going too.
In such a natural-language system, a user can often avoid the need to locate a report or configure filters. Instead, a user could ask, “what were our receivables at month end?” or tell the system “compare this quarter’s revenue with the figures from the same time last year.” The system interprets the request, pulls the data from the ERP and returns an answer or a generated report.
“Your ERP ends up becoming a system of record.”
Joe Locandro
Cloud-based enterprise software provider Workday is also pushing conversational reporting, under the slogan “AI is the new UI”. The point is not just to use a chatbot to help users click around an ERP. The point is to use an AI to generate on the spot the reporting views the user needs.
In such a system, the software provides a series of AI-generated views – pages, charts and summaries assembled on demand. These give finance teams a faster explanation, without forcing them into predetermined screens.
The ERP vendors are also experimenting with tools that let business users automate work without specialist help. At SuiteWorld, Oracle NetSuite’s biggest annual user event, the company highlighted new AI tools including Prompt Studio for designing and testing prompts. It also showed workflow-related assistants aimed at making automation easier to build and govern.
Xero and QuickBooks already offer rules and bank-feed automation. If enterprise ERPs are a reliable signal, the next wave in accounting software will be less about prettier dashboards and more about making reporting and automation across the wider finance stack dramatically easier.
Where will the productivity gains appear?
Wading through all the promises that vendors are now making about generative AI, it is worth asking where the first real productivity gains will appear.
Right now, the most credible gains are in reporting: producing management commentary, variance explanations, board-ready summaries and the like from plain-English queries, without hours spent pulling data into spreadsheets and writing the story.
But reporting is only really useful if the data it relies on is accurate and properly reconciled. That is why the bigger opportunity is task automation. If software can reliably reconcile employee expenses, run payroll and lodge returns, accounting teams and firms may see productivity lift substantially.
There is little evidence we are anywhere close to that. Vendors have shared few details on which tasks will be automated first, how accuracy will be validated, or how they will manage known risks such as hallucinations and errors.
There is also a third, more immediate productivity benefit: if ERP software becomes usable by business owners and frontline managers, the reporting and scenario work stops being trapped inside specialist teams.
ERPs can bury valuable information several menus deep. A conversational interface could let frontline managers ask for inventory cover, margin erosion or customer profitability without hunting through screens or configuring reports, making the system more accessible across the business.
That would remove one of the biggest barriers to ERP adoption: complexity. If smaller companies can access ERP insights through a chatbot, the constraint becomes less about usability and more about price.
This is where small-business accounting platforms should be paying attention. In April 2025, I asked Evan Goldberg, NetSuite’s founder and executive vice-president, whether Oracle would push further into smaller businesses if the main interface became conversational. He said it was something the company was seriously considering. “More to come,” he said with a smile.
Where will the AI agent live?
There is another possibility: we stop using finance software directly altogether. If information is mediated through a chatbot, the next question is where that chatbot lives. Does it sit inside a specific product, such as Xero, or does it float above everything, pulling data from whatever systems the business uses?
Rimini Street, which provides third-party support services for enterprise software including Oracle and SAP, argues the second model will work better. It says an external agent that can read across accounting, inventory management and CRM tools will produce better results than separate in-app agents trying to talk to each other.
That would turn accounting and ERP platforms into “headless” systems of record: users would interact with the agent, while the underlying applications acted as compliant stores of structured data.
“Your ERP ends up becoming a system of record,” says Joe Locandro, Rimini Street’s EVP and global CIO. “The functionality that’s left in an SAP or a Salesforce or whatever is pretty well transactional records, and the data is being pulled from them with agents through the data fabric.”
If the interface is separated from the system of record, the agent can also draw from multiple databases at once. That could make cross-system reporting far easier, ending the need to copy finance, operational and customer data into a single warehouse first.
In December 2025, Rimini Street launched Rimini Agentic UX, an AI-driven engagement layer it says sits on top of existing ERP systems to automate work.
An agent future?
For accountants who live in Xero, MYOB or QuickBooks all day, the most disruptive possibility is this: you may not need to be in those programs as often. If conversational tools keep improving, the day-to-day experience could shift from operating software to supervising an agent that operates it for you.
That creates a fork. We could end up with tighter suites and more consolidation among accounting and ERP vendors. Accounting software could gain ERP-like capability behind the scenes in the platforms we already use.
Or we could end up with independent personal agents working across the finance stack and beyond, with the accounting system reduced to a compliant system of record.
Whichever path wins, the direction is the same: AI is shifting value away from reconciliation and compliance and toward the layer that explains performance and supports decisions. As the system of record becomes more automated, the real contest is who owns reporting and forecasting – and who becomes the trusted agent that turns financial data into a narrative you can act on.
