Acumatica 2026 R2 Brings AI Agents and External AI Tools Under ERP Controls
Acumatica has made 2026 R2 generally available, introducing a significant set of updates to its cloud ERP platform for mid-market businesses.
Released on 1 October, the latest version places a strong emphasis on embedded AI, analytics, user experience and industry-specific capabilities. More importantly, it gives businesses greater control over how AI interacts with ERP data and business processes.
The release introduces AI agents that can respond to business events, work with other agents and automated processes, as well as new Model Context Protocol (MCP) servers that allow external AI tools such as ChatGPT, Claude and Gemini to access selected ERP data.
Acumatica is also making its Modern UI the default experience, while preparing to retire Classic-only functionality with the planned 2027 R1 release.
AI Agents Move from Experimentation to Business Automation
One of the biggest changes in Acumatica 2026 R2 is the introduction of AI agents designed to automate practical business processes.
Users can create agents using plain-language instructions and save them for repeated use. Agents can be triggered by a Business Event, Import/Export Scenario or existing workflow.
They can then use information from Generic Inquiries, Acumatica’s configurable data-query framework, to understand the relevant business context and take defined actions on eligible records.
Agents can also pass work to another AI agent or an automated process, allowing businesses to build multi-step workflows.
Potential applications include:
- Adjusting prices to maintain target margins
- Qualifying sales leads
- Assessing customer credit risk
- Automating routine operational decisions
- Passing completed tasks to the next stage of a workflow
This moves AI beyond simple question-and-answer functionality and towards controlled, process-driven automation inside the ERP system.
AI Assistant Goes Deeper into Financial Analysis
Acumatica’s AI Assistant is also gaining stronger financial analysis capabilities.
It can examine document history to explain how an account balance has changed over previous periods or at specific points in time. Users can also save AI-generated answers as dashboard widgets and review the underlying inquiry used to produce the result.
That provides an important level of transparency for finance teams. Rather than simply receiving an AI-generated answer, users can trace the response back to the ERP data and inquiry behind it.
Security and Governance Remain at the Centre
As AI takes on more responsibility, controlling what it can access becomes increasingly important.
Acumatica 2026 R2 keeps AI agents within the ERP’s existing security framework. Agents operate using the access rights already assigned to the relevant user.
Administrators can also mask selected fields before information is sent to a large language model. The original values can then be restored within Acumatica.
The release adds further visibility through processing logs, including provider-reported token usage for AI automations. Administrators can also configure warnings when AI-unit consumption reaches defined thresholds.
For organisations deploying AI at scale, these controls can help provide a clearer balance between automation, security and cost management.
External AI Tools Can Now Access Selected ERP Data
Another important development is the introduction of Model Context Protocol (MCP) servers.
MCP allows compatible AI assistants, including ChatGPT, Claude and Gemini, to interact with selected Acumatica data without requiring users to export information manually.
Administrators can expose published Generic Inquiries as callable tools. External AI assistants can then use those inquiries to access approved ERP information on a read-only basis.
Importantly, requests continue to operate under the user’s existing access permissions and are logged.
This means organisations can give employees a more natural way to ask questions about ERP data while retaining control over which information is exposed.
A separate DAC-level security setting also limits the underlying data objects that business users can query when creating their own inquiries.
A More Scalable Data Foundation for Analytics
Acumatica is also changing how reporting and analytics interact with the ERP database.
The new release uses a dedicated, performance-optimised copy of ERP data for reporting and analytics. The aim is to prevent demanding reports from competing with operational processes such as order entry and invoicing.
Acumatica says this native data warehouse will form the basis for further enhancements in future releases.
The company has also introduced Generic Inquiry Optimisation, which can store the results of large or frequently used inquiries and refresh them either on a schedule or when required.
Those results can then be reused across:
- Dashboards
- OData feeds
- Pivot views
- MCP connections
For businesses with large volumes of ERP data, this approach could improve analytical performance while reducing the impact on day-to-day transactional workloads.
Industry-Specific Enhancements
The 2026 R2 release also introduces new capabilities across Acumatica’s industry editions.
Construction
Construction businesses gain several new tools, including:
- Parent-child projects
- S-curve cost forecasting
- A new vendor portal
Acumatica describes the vendor portal as the first step towards a broader Construction Exchange.
Manufacturing
Manufacturers receive a new Advanced Planning and Scheduling (APS) engine designed to improve production planning and scheduling capabilities.
Distribution
Distribution businesses can now see Available-to-Promise dates directly on order lines, providing clearer information about when products can be fulfilled.
Acumatica also reports performance improvements, with payroll calculations and batch releases running up to 63% faster.
Modern UI Becomes the Default
The Modern UI is now the default user experience in Acumatica 2026 R2.
Classic screens and functionality have not disappeared immediately. However, Acumatica plans to retire Classic-only forms, dashboards, pivot tables and customisations with the 2027 R1 release.
Businesses that still rely heavily on Classic functionality should therefore begin assessing their current environment rather than waiting until the retirement deadline approaches.
What Acumatica 2026 R2 Means for ERP Teams
The latest release represents a broader shift in how AI is being incorporated into ERP systems. Rather than treating AI as a separate tool layered on top of business applications, Acumatica is bringing AI agents and external AI access into the same security and data structures that already govern the ERP.
1. Treat the Generic Inquiry layer as an AI boundary
Generic Inquiries now play an important role across AI agents, AI Assistant and MCP connections.
ERP teams should review which inquiries are published, which roles can access them and what information they expose. This provides a practical starting point for determining what AI tools can see.
2. Define triggers before deploying AI actions
AI agents can respond to business events and pass tasks between different agents or automated processes.
Before deploying these workflows, organisations should map out:
- What triggers each agent
- What actions it can take
- Which records it can modify
- Where human approval is required
- How the workflow can be audited
This becomes particularly important as businesses move from individual AI tasks towards multi-step automation.
3. Start planning the Classic UI migration
The retirement of Classic-only functionality in 2027 means organisations should begin identifying affected screens, dashboards, pivot tables and customisations during their 2026 R2 planning cycle.
Testing replacements early will give ERP teams more time to address compatibility issues and prepare users for the transition.
The Bigger Picture
Acumatica 2026 R2 is more than a collection of new AI features. It represents a move towards controlled AI automation within the ERP environment.
AI agents can act on business events, external AI assistants can access approved ERP information, and both are tied back to existing permissions, inquiries and audit controls.
For ERP teams, the key question is no longer simply whether to use AI. It is increasingly about where AI should have access, what it should be allowed to do and how those actions can be governed.
With 2026 R2, Acumatica is positioning its ERP platform to make that transition while keeping data access, security and business processes within the organisation’s existing controls.