Automation environments grow faster than the teams that run them. Jobs multiply across SQL Server, Azure Data Factory, Airflow, SAP, JDE, and Banner, and when one fails, finding the root cause often means searching several consoles at once, frequently outside business hours. At the same time, IT leaders carry pressure to adopt AI while staying accountable for where operational data goes. JAX and JAMS MCP close both gaps together.
JAX is an AI agent that runs inside the JAMS Web Client. It finds jobs, troubleshoots failures, and answers how-to questions in plain language, with each response grounded in the JAMS user guide and checked against a built-in glossary. JAX acts only when a user asks it to. Reads flow freely, and every write action pauses for the user’s explicit approval before it runs. JAX does not learn between sessions, and conversations are not retained on the server.
JAMS MCP is a connector, built on the open Model Context Protocol standard, that brings JAMS into the AI tools engineering teams already use, including Cursor, VS Code with Copilot, Claude Code, Claude Desktop, and Codex. Users query jobs, investigate failures, and manage runs in plain language without leaving their tool.
Both capabilities run inside the customer’s own network and act as the signed-in user, with that user’s exact JAMS permissions. There is no elevated AI account: whatever a user cannot do in the JAMS interface, JAX and JAMS MCP cannot do on that user’s behalf.
Every JAX and MCP operation is recorded in its own dedicated log, and changes made through the JAMS API land in the JAMS audit trail like any other change. Customers choose their own AI model, whether a commercial provider such as OpenAI or Anthropic or a model running entirely on their own hardware, and JAMS never trains on customer data. In the current release, neither feature edits or deletes a job, folder, schedule, or agent definition.
“Our customers have told us for years that the slowest part of their day is not the automation itself, it is finding out why a job failed,” said Pete Hegland, Chief Executive Officer of JAMS Software. “JAX and JAMS MCP answer that question in plain language, and they do it without asking a single customer to hand their job scripts and logs to a system they do not control.”
“We designed JAX and JAMS MCP around one principle: the AI can never do more than the person using it,” said Dan Fowle, Product Manager at JAMS Software. “Every action runs as the signed-in user with that user’s exact permissions, every write waits for confirmation, and customers can run the model entirely inside their own network. We will add capabilities over time, and that control model will stay exactly the same.”
JAX and JAMS MCP are available now to all JAMS Web customers, with no separate license, SKU, or additional cost. AI-assisted creation of new jobs and workflows from a plain-language description is on the roadmap for a future release, gated by the same approvals and permissions as every other action.
