Your Accounting Work May Be Ready for a Better Workflow
If you run a Malaysian SME, your finance work may depend on spreadsheets, email attachments, WhatsApp messages, bank downloads and an accounting system that still needs substantial manual input. Your team may spend hours matching payments, checking invoices, preparing reports and answering questions that should be easy to resolve.
The problem is not simply that accounting takes time. Manual work can delay decisions, hide errors and make it harder for you to see what is happening across sales, purchasing, payroll and cash collection. When your business grows, the same process becomes more difficult to control because more transactions pass through more people.
An emerging category of AI-native accounting software is designed around automated workflows and AI agents rather than treating AI as an extra feature added to an old system. Rillet, the company covered in the source article, raised US$100 million at a US$1 billion valuation after showing that businesses were willing to replace established accounting and ERP systems with an AI-focused alternative. Source: TechCrunch
TL;DR
AI accounting is moving beyond simple data entry towards multi-step finance workflows, with humans reviewing important decisions. For your SME, the practical lesson is to improve data quality, automate repetitive processes and require clear approval controls before adopting new tools.
What This Means
Traditional accounting software usually stores records and helps you produce reports. You or your staff still need to enter information, match documents, check unusual transactions and move data between systems. AI-native accounting platforms are built to let software agents carry out parts of that work, such as reading an invoice, identifying the supplier, matching it with a purchase order, posting the transaction and sending it for approval.
The important distinction is assistance versus independent action. A basic AI feature might suggest an account code. A more capable agent may complete several connected steps across a workflow. That can reduce repetitive work, but it also creates a new requirement: you must be able to see what the system did, which data it used and who approved the result.
Rillet’s product includes a governance feature that allows accountants to audit decisions made by its AI agents, including the numbers used and how calculations were performed. The company also says its system prevents customer data from being used to train models across customers and allows model routing to selected foundation models. Source: TechCrunch
This matters because financial records are not ordinary business data. They contain supplier details, customer information, employee records, bank activity and commercially sensitive information. Automation should make your controls stronger, not make decisions invisible.
How This Applies to Malaysian SMEs
For a trading or distribution business, AI can help connect purchase orders, delivery documents, supplier invoices and payment records. If your staff currently download documents from email and manually compare them, an automated workflow could flag mismatched quantities, duplicated invoices or invoices received without an approved purchase order. You would still decide how exceptions are handled, but your team would spend less time searching for routine differences.
For a services company, the useful application may be invoice preparation, expense classification, project tracking and receivables follow-up. A system could identify billable work from approved records, prepare a draft invoice and alert you when a customer account needs attention. This is especially useful when project managers, sales staff and finance personnel keep information in separate places.
For a restaurant, retailer or multi-outlet operator, automation can help bring together daily sales, payment channels, inventory movements and supplier transactions. Instead of waiting until month-end to discover that one outlet’s figures do not match, you could receive exception alerts during the week. The value is not just faster bookkeeping; it is earlier visibility into operational problems.
For a Malaysian SME working with an external accountant, AI does not necessarily remove the need for professional advice. It may improve the quality and timing of the information your accountant receives. Your accountant can spend more time reviewing unusual transactions, preparing management insights and supporting compliance, while routine document handling is managed through a controlled workflow.
You should also consider local requirements before selecting a platform. Your process may involve Malaysian tax records, e-Invoice readiness, payroll information, foreign-currency transactions, audit documentation and personal data obligations. A tool that works well in another market may still require integration, configuration or review before it fits your business.
Numbers Behind the Trend
| Data point | What it suggests for you |
|---|---|
| US$100 million raised by Rillet | Investors are showing strong interest in AI-focused finance software. |
| US$1 billion valuation for Rillet | The market sees accounting as a wider finance workflow opportunity, not only bookkeeping. |
| 600 Rillet customers | Businesses are testing alternatives to established accounting and ERP systems. |
| 50% of Rillet customers reportedly came from Intuit | Some users are replacing existing tools rather than adding another system beside them. |
| 30% reportedly came from NetSuite and Sage Intacct | AI-native tools can also appeal to businesses that already use established cloud platforms. |
| 20% reportedly came from Oracle, SAP, Workday and Microsoft products | Switching pressure can extend into larger business software environments. |
| At least 5% projected growth in accounting-related needs through 2034 | Automation is more likely to reshape accounting work than remove the need for finance expertise. |
All figures in this table are reported by TechCrunch from its interview and cited sources. Source: TechCrunch
Practical Takeaways
- Map one workflow first. Choose a process such as supplier invoice approval, expense claims or payment reconciliation. Do not begin by trying to automate your whole finance function.
- Measure the current process. Record how many documents are handled, how long approval takes and how often corrections are needed. This gives you a useful baseline.
- Clean your master data. Standardise supplier names, customer records, chart-of-accounts categories and approval limits before introducing automation.
- Separate suggestions from approvals. Let software prepare or classify transactions, but require an authorised person to approve sensitive actions.
- Demand an audit trail. You should be able to see the original document, the system’s interpretation, any calculations, changes made and the final approver.
- Check data handling. Ask where information is stored, whether customer data is used for model training, how access is controlled and what happens when you export your records.
- Review Malaysian compatibility. Confirm that the system supports your accounting structure, tax workflow, reporting needs and integrations.
- Train staff around exceptions. Your team should know how to review unusual transactions instead of accepting every automated suggestion.
- Keep a fallback process. Document what happens if the AI service is unavailable, produces an uncertain result or cannot read a document.
The best first use of AI in accounting is not “let the system do everything.” It is “let the system handle repeatable steps while your people focus on judgement, exceptions and decisions.”
The Bigger Picture
The longer-term change is that accounting software may become less like a digital filing cabinet and more like an operational assistant. Instead of opening several systems to find an answer, you may ask a finance agent to explain why receivables changed, identify overdue approvals or prepare a report based on authorised records.
That does not mean every SME should replace its accounting platform immediately. Established systems may still be suitable, especially when your data is reliable and your workflows are already efficient. The more practical question is where manual work is slowing you down and whether a controlled automation layer can improve that process.
The source article also highlights that public-company rules require human approval for transactions made by AI agents. Source: TechCrunch While your SME may not face the same governance requirements, the principle is sensible: financial automation needs accountability. You should know which actions can happen automatically, which require review and how decisions can be traced later.
For you as a business owner, the opportunity is to build a finance function that gives you clearer information without adding unnecessary administrative work. Start with one high-volume process, protect sensitive data, keep human oversight and judge the result by whether you receive accurate information sooner. That approach will help you adopt AI because it solves a real business problem, not simply because the technology is receiving attention.
Ready to Streamline Your Operations?
Your business should run itself. AutoRunBiz deploys AI agents to automate your daily operations — WhatsApp orders, invoicing, customer follow-ups, and accounting. Book a free 15-min ops audit to see where automation fits your business →