2026 Virtual DevCon Schedule

Welcome to Acumatica Virtual DevCon 2026 and the theme for the day: making AI enablement easier for partners, customers, and prospects.

Overview of the Beginner → Intermediate → Advanced path.

Where AI is headed on the platform: what’s next for AI Studio, AI agents, and external AI connectivity, and what developers and partners should build now.

• Understand the platform’s AI direction
• Identify opportunities for partners and developers

ㅤ

A complete tour of Acumatica’s AI product line and how the pieces fit together: AI Studio and AI Automation, AI Assistant (widgets, prior-period analysis, Help Agent), AI agents and agentic workflows, MCP servers, and AI embedded in the modules (anomaly detection, intelligent bank matching, CRM summaries and prioritization, ML cross-sell/upsell). Covers what attendees need before turning anything on: version and Modern UI requirements, licensing, AI Units, cost controls, the secure cloud proxy architecture, and LLM basics (prompts, system instructions, tokens, model choice).

• Name every AI feature in 2026 R2 and the problem it solves
• Know the prerequisites, licensing, and AI Unit model
• Explain how requests flow from Acumatica to an LLM provider
• Know which features are native vs. bring-your-own LLM

A live, end-to-end walkthrough from a clean 2026 R2 instance. Configure: enable the features, create and test an LLM connection, set up data masking, access rights, and reusable system instructions. Use: run a built-in prompt, adapt one from the AI Automation Prompt Library, and build a custom prompt. Then enable AI Assistant for users, ask natural-language data questions, save an answer as a dashboard widget, use the Help Agent, and check statistics and token usage.

• Enable AI features and create/test an LLM connection
• Apply data masking, access rights, and system instructions
• Use built-in, library, and custom prompts
• Roll out and use AI Assistant, widgets, and the Help Agent
• Monitor statistics, tokens, and AI Units

ㅤ

For developers and technical consultants. Build an AI agent with plain-language instructions, Generic Inquiry data sources, and tool calling; trigger it from a Business Event and hand work to a second agent. Then configure an MCP server that exposes published Generic Inquiries as read-only tools, connect an external AI client (Claude, ChatGPT), and cover governance: access rights, DAC-level security, Generic Inquiry Optimization, token usage, and when to use MCP vs. REST/OData/webhooks.

• Create an agent with instructions, GI data sources, and tool calling
• Trigger agents from Business Events and chain them
• Configure an MCP server and connect an external AI client
• Apply governance and choose MCP vs. API patterns