Agentic and MCP ready
An LMS your AI agents can work with, safely.
LaserLMS exposes a Model Context Protocol (MCP) server and a scoped API, so AI assistants can answer training questions from live data and take the actions you approve.
// Example client configuration
{
"mcpServers": {
"laserlms": {
"url": "https://acme.laserlms.com/mcp",
"headers": { "Authorization": "Bearer ${LASERLMS_TOKEN}" }
}
}
}Questions answered from live training data.
Managers ask in plain language, in the assistant they already use. The assistant calls LaserLMS, gets the facts and proposes the next step.
- No report hunting. Ask who is overdue and get names, not a link.
- No copy and paste. The assistant reads LaserLMS directly.
- No surprises. Any change is proposed first and runs only when approved.
Who at the Northgate warehouse is still outstanding on Workplace Health and Safety?
get_compliance_status(course: "Workplace Health & Safety", site: "Northgate", status: "outstanding")
Seven people are outstanding. Four have not started and three are more than halfway through. Shall I draft a reminder to all seven for your approval?
Yes, and copy their shift leads.
send_reminder(users: 7, cc: "managers", template: "overdue")
Example conversation.
Example tools an agent can use
The tool set is configured per deployment, and each tool respects the permissions of the token that calls it.
| Tool | Access | What it does |
|---|---|---|
| get_compliance_status | Read | Compliance status by course, company, department, city or person |
| list_overdue_learners | Read | People past their due date, with progress and manager |
| get_user_progress | Read | One person's courses, status, scores and certificates |
| search_courses | Read | Find courses by title, category or keyword |
| enrol_users | Write, approval | Enrol people or cohorts in a course with a due date |
| send_reminder | Write, approval | Send a reminder using an approved template |
| create_course_from_brief | Write, approval | Ask CourseTitan to draft a hidden course for review |
Guardrails come first
Agents are only as safe as the permissions behind them. LaserLMS treats an agent like any other integration, with tighter defaults.
Scoped tokens
Each token is tied to a company and role, so an agent never sees more than the person it acts for.
Read-only by default
Write scopes are opt-in, per tool, and can require human approval for every call.
Full audit trail
Every agent call is logged with the tool, the inputs, the result and who approved it. Revoke access at any time.
What teams automate
New starter arrives
An HR agent sees the new record and enrols the person in their role's onboarding path.
Weekly compliance digest
An agent summarises what is outstanding per site and drafts reminders for approval.
Policy changes
A drafting agent asks CourseTitan for a refresher course and assigns it once a reviewer signs off.
Manager questions
Managers ask about their team in chat instead of opening reports.
Related capabilities
Agent and MCP questions
Something else? Ask us in a demo.
What is the Model Context Protocol (MCP)?
MCP is an open standard that lets AI assistants connect to tools and data in a consistent way. An MCP server tells the assistant which actions are available, and the assistant calls them when they help answer a request.
Which AI assistants work with LaserLMS?
Any MCP-compatible assistant or agent platform, including Claude, ChatGPT and Microsoft Copilot Studio, as well as agents your developers build. Platforms without MCP can use the REST API.
Can an AI agent change data in LaserLMS?
Only if you allow it. Tokens are read-only by default. Write actions such as enrolling users or sending reminders need explicit scopes, and can require a person to approve each action.
Does the MCP server work with on-premises deployments?
Yes. The MCP server runs alongside your LaserLMS instance, so on-premises customers keep agent traffic inside their own network.
Connect your assistant to LaserLMS.
See an AI assistant answer real training questions against a LaserLMS demo company, with approvals switched on.
- Tailored to your use case, not a slide deck
- Cloud, dedicated or on-premises options explained
- Clear next steps and pricing for your scale