How Dynamics 365 Business Central 2026 AI Agents Transform Operations for Australian SMBs

The Shift From AI Assistance to AI Operation

Australian small and midsize businesses have been hearing about AI in their ERP systems for years. Most of what they have seen so far has been assistive: Copilot suggests a vendor match, a chatbot answers a question, a dashboard surfaces a trend. Useful, but still fundamentally a tool that waits for you to ask.

The 2026 Release Wave 1 for Dynamics 365 Business Central changes that. For the first time, Business Central includes autonomous agents that take action on real business processes. They read incoming invoices. They match vendors to general ledger accounts based on your historical patterns. They prepare approval workflows. They process sales orders from receipt through confirmation. All of this happens with human oversight at key decision points, and every action is visible on the document record with a clear "Created by AI" indicator.

This is not a roadmap promise. These agents reached general availability and are rolling out to Australian tenants now, included in your existing Business Central license at no extra cost.

As a Microsoft AI Cloud Partner headquartered in Sydney, Oxana has been tracking this release closely. Through our Microsoft Dynamics 365 services, we are helping Australian organisations across Sydney, Melbourne, and Brisbane understand what these agents do, where to start, and how to deploy them safely.

The Three Autonomous Agents You Need to Know About

1. Payables Agent

The Payables Agent is the headline feature of this release wave, and it is where most finance teams should start.

What it does:

• Reads incoming supplier invoices automatically
• Matches each invoice to the correct vendor and general ledger account based on your historical coding patterns
• Prepares the approval workflow so your finance team reviews rather than manually entering data
• Categorizes processed emails in shared inboxes so nothing is duplicated or missed

Why this matters for your finance team: Most Australian SMBs with moderate transaction volume spend 15 to 30 hours per week on manual invoice processing. The Payables Agent does not eliminate the finance function. It shifts the team from data entry to review and exception handling, which is where their expertise actually adds value.

2. Sales Order Agent

The Sales Order Agent handles the full sales order lifecycle without manual intervention at each step.

What it does:

• Receives sales orders through connected channels
• Validates and processes the order through to confirmation
• Reduces manual coordination in high volume order environments
• Flags exceptions for human review at defined checkpoints

Why this matters for operations leaders: If your team processes more than 50 orders per day, the manual touchpoints between receipt, validation, and confirmation create bottlenecks. The Sales Order Agent removes those bottlenecks while keeping a human in the loop for exceptions.

3. Custom AL Agents

This is where Business Central becomes a platform rather than just a product. Microsoft now allows developers to build custom agents using AL code (the development language for Business Central extensions), triggered by your own business events.

What this means: If the Payables Agent and Sales Order Agent do not cover a process unique to your industry, your implementation partner can build a custom agent that does. Think of the default agents as templates. AL extensions let you adapt them to your specific workflows, whether that is manufacturing job costing, project billing, or industry specific compliance requirements.

Oxana's expertise and credentials, as both a Microsoft AI Cloud Partner and an Anthropic Partner, bring a unique combination of Microsoft Business Applications expertise and applied AI engineering to these custom agent projects, alongside our AI driven website solutions. We can help you assess which processes are ripe for custom automation and build agents that fit your operations precisely.

Copilot: Smarter, More Extensible, Better Connected

Alongside the autonomous agents, the 2026 Release Wave 1 significantly upgrades Copilot within Business Central.

Model Context Protocol (MCP) Support

Available now in version 27.1 and later, MCP enables Copilot Studio agents to interact directly with Business Central data. This removes one of the biggest barriers to building cross system AI workflows. If you have been waiting to build agents that span Business Central and other systems, MCP is the bridge you needed.

Automated Report Generation

You can now run reports across all dimension values automatically, or combine multiple reports into a single PDF. For finance teams closing the books each month, this eliminates hours of manual report assembly work.

Centralized Outlook Task Pane

All agent generated tasks are visible in a single pane within shared Outlook inboxes. Nothing gets missed or duplicated between human and AI processing. This is a small feature with a big operational impact: it gives your team a single view of everything the AI has done and everything that needs human attention.

Copilot Extensibility

Developers can now build custom Copilot skills for scenarios that Microsoft has not covered out of the box. If your team needs Copilot to answer a question specific to your business, your partner can build that skill.

Finance, Supply Chain, and Power Platform Updates

Beyond the AI agents and Copilot upgrades, Release Wave 1 delivers practical improvements across finance operations, supply chain, and the Power Platform.

Finance:

• AI assisted bookkeeping handles routine data entry and categorization based on transaction history
• Subscription revenue analytics fills a long standing gap in native reporting for businesses with recurring revenue
• Enhanced international tax compliance tools for global operations

Supply Chain:

• AI driven warehouse picking creates intelligent picking sequences to reduce travel time and error rates
• Inventory rebalancing provides automated suggestions for redistributing stock across locations based on demand and lead times
• Expanded electronic document exchange for compliance and faster processing

Power Platform:

• Power Automate agent features include self healing flows that detect and correct their own failures, reducing maintenance burden
• Power Apps generative pages create AI generated app pages that surface Business Central data without full development

Comparison: Business Central Before and After 2026 Release Wave 1

How to Prepare Your Organisation

Enable Now

Start with the Payables Agent in a sandbox environment. Test it against your real invoice data to see how it handles your vendor patterns and GL coding conventions. Enable the AI transparency indicators immediately so your team builds confidence seeing what the agents do and how they do it. If you are building custom agents, enable MCP support in version 27.1 or later.

Plan This Month

Review the Sales Order Agent for your order volume. If you process more than 50 orders per day, this agent will deliver measurable time savings. Evaluate automated report generation for your most manual monthly reports. If your business has recurring revenue, review the subscription revenue analytics app.

Test Before Production

Custom AL agents and Copilot extensions require sandbox validation against your real data before enabling for all users. This is not optional. Autonomous agents that take real business actions need to be tested with real business data before they go live.

Build Team Confidence

The transparency controls in this release are designed to help organisations expand agent scope gradually. Start with one agent in one process. Let your team see the "Created by AI" indicators. Let them use the instant stop control during testing. Confidence in AI agents comes from transparency and control, not from mandates.

Australian Regulatory Context

Privacy Act 1988: If your invoice processing or sales order data includes personal information, ensure your AI agent configuration aligns with your obligations under the Australian Privacy Principles. The agents process data within your existing Business Central tenant, which helps with data residency, but you should document the agent activity as part of your privacy impact assessment.

ASIC and ATO Compliance: Automated invoice processing and bookkeeping must still meet your obligations under Australian accounting standards and ATO record keeping requirements. The "Created by AI" indicators and audit trail in Business Central support compliance, but your finance team should review the agent output as part of normal month end procedures.

Essential Eight Considerations: Autonomous AI agents that take business action introduce new considerations for application control and restricted execution. Oxana, as an SMB1001 Gold Certified provider, can help you assess the security implications of enabling AI agents in your Business Central environment.