- Key insight: Learn why Acumatica is routing reporting queries through a native data warehouse.
- What's at stake: Organizations relying on classic dashboards slated for future removal.
- Expert quote: "Growing businesses aren't looking for more AI experiments..." — Jon Pollock, Acumatica
Mid-market business ERP provider Acumatica announced the rollout of its 2026 R2 release, featuring enhanced AI-driven capacities for automation, analytics and integrations, as well as user experience changes, such as the retirement of the classic interface.
"Growing businesses aren't looking for more AI experiments; they're looking for practical AI solutions that empower teams to work smarter every day," said Jon Pollock, chief product officer at Acumatica, in a statement. "With 2026 R2, we've embedded intelligent capabilities throughout the platform while continuing to simplify the user experience and deliver deeper innovations tailored to the specific industries our customers serve."
Many of the AI-related upgrades can be rooted in two new additions to Acumatica: Model Context Protocol support and a native data warehouse.
Users can now configure MCP servers and expose published generic inquiries as callable tools without the need to export, upload or duplicate information in a separate AI workspace, thus reducing the need for specific integrations. Instead of building a separate integration for each client, users can now configure one or more MCP servers in a tenant and add a focused set of tools to each server. For example, an organization might connect Acumatica ERP to AI assistants, such as Claude or ChatGPT and work with their data through there. Role-based permissions, inherited security and field-level data masking automatically carry over, allowing teams to use the AI tools they already rely on while IT retains centralized governance and auditability for every query.
The native data warehouse, meanwhile, means that reporting and analytics now run against a dedicated, performance-optimized copy of ERP data instead of the live transactional database, so reports don't compete for system resources with day-to-day order entry, invoicing and approvals.
Every time a generic inquiry (a built-in reporting tool that lets people pull and display data from their database tables without writing code) is opened, the system does not query and prepare the data set of the generic inquiry according to all its settings, but instead loads the ready data set from the data warehouse. As a result, the generic inquiry is opened faster. When a generic inquiry is configured to store its data set in the data warehouse, dashboards, OData, MCP, pivot view and grouped view also retrieve the stored data instead of executing the inquiry query directly.
The release also comes with a host of new enhancements for its AI Assistant. This includes finer options for widget control, such as determining whether the assistant includes widgets in its answers at all, the ability to generate its own widgets or further refine existing widgets in follow-up prompts or add them to the dashboard. The assistant has also been upgraded to provide technical and user support: People can get answers on how to complete a task, find functionality or troubleshoot an issue using the same chat used for data questions; such help-related questions do not consume AI units. Users can combine questions about business data and help in a single prompt.
The AI also supports new automation capacities. Among them is the ability for agents, via a new "tool calling" ability, to retrieve specific records and apply line-by-line updates, which enables finer-grade automations for increasing unit prices across many lines (either all lines or only those that meet certain conditions, such as belonging to a warehouse), applying discounts to selected lines, and adding a free item when an order total exceeds a threshold.
Automation capabilities can also be brought directly into business processes to build agentic workflows by combining the AI automation agent with Acumatica ERP mechanisms. For example, agents can review records, retrieve supporting data, write conclusions to the current record, and pass those results to other configured steps. Agents can use the current record and supporting data from generic inquiries.
The tool-calling ability also helps keep data private. Previously, users protected sensitive data by inserting masked field placeholders directly into AI Automation instructions. Because data now flows through the tools agents use, versus the LLM instructions themselves, AI Automation replaces the values of these fields with placeholders before sending data to the LLM. When AI Automation receives a response, it restores the real values before saving the result or showing it in the system.
There are also new features meant to rein in AI spend. The AI assistant, for example, can now display a warning when AI unit usage reaches the configured threshold so that the user does not exceed token allotments. Users can also review the Acumatica ERP token counts for each request, compare the Acumatica ERP and provider-reported token consumption totals for the entire execution, understand differences in provider totals (which vary by provider and model and may include provider-specific categories), and review the provider's input, output and total token counts.
Another new feature allows professional services businesses, such as accounting firms, to analyze project revenue for a broader range of projects. People can analyze projects regardless of the billing rules assigned to project tasks, and use updated revenue calculations and aging buckets to do so. Users can also move eligible expenses between projects and project tasks without manually reversing and recreating transactions in order to keep project budgets aligned with actual project expenses.
The latest release also features a number of new payment options through its integration with AP automation solutions provider AvidXChange, tap-to-pay functionality via iPhone integration, and enhanced ACH clearance and international payments via integration with Bill. There's also a host of new industry-specific features and capacities focused on construction, e-commerce, manufacturing and distribution.
The new version also has also retired the classic user interface in favor of a Modern UI that is now the default user interface. Users can no longer switch between it and the Classic UI. Screens that are available only in the Classic UI can still be opened in 2026 R2, but they display a warning that they will be deprecated in the next release. When the Classic UI is disabled, creating Classic dashboards or pivot tables is no longer possible. Existing Classic dashboards and pivot tables can still be accessed in 2026 R2, giving organizations an opportunity to migrate the ones they need before the Classic UI functionality is removed in 2027 R1.
"We're focused on making advanced AI capabilities easier to apply in the real world," said Miten Mehta, chief engineering officer at Acumatica, in a statement. "With 2026 R2, we enable our customers and partners to build agents with plain text prompts and generic inquiries. We are giving teams more intuitive ways to access insights, automate routine work and adapt the platform to how they operate."
The 2026 R2 release is now generally available for Acumatica users.
Introductory bullet points created by AI with editorial review.
