Welcome to Trimble Assist
Trimble Assist is the conversational AI application for finding and working with agents across the Trimble construction portfolio.
People use Trimble Assist in SketchUp, Tekla, Trimble Connect, ProjectSight, Vista, and other Trimble products, and in the standalone web app at assist.ai.trimble.com. The chat, the conversation history, and the sign-in stay the same from one product to the next.
See the following sections to learn more.
How Assist is delivered
| Form | When to use it |
|---|---|
| Standalone web app | Deep research, multi-monitor work, or anything outside a specific product |
| Embedded widget | Help inside a product, in a panel that feels native to that application |
You integrate into Assist, not around it. Assist takes care of the layouts, conversation history, sign-in, and enterprise compliance for the chat (aligned with ISO 27001, ISO 42001, and SOC 2 Type 1, whose third-party audits have been passed with official issuance expected in November 2026, and with GDPR). Your team owns the content that makes Assist valuable in your domain.
More than a chat sidecar
Assist scales from a narrow side panel to a full-screen workspace. With MCP Apps, an agent can show what the task needs inside the conversation: an interactive RFI builder, a clash dashboard, a procurement comparison, a model viewer, or a form.
That turns Assist from something that answers questions about the current page into a way to deliver applications. Agents can demonstrate product value, guide people through workflows, and offer interactive tools without sending anyone out of the conversation.
Design agents for a portfolio audience
Because Assist is embedded across many products, people may find your agent while they are outside your application. The agent should check where the person is, explain what it can do once the right product or license is available, and point them to the next step (a tutorial, a trial, or the right application) instead of failing on a tool that only works inside your product.
Pass run context from the embedding application so the agent's instructions can adapt to where the person is working.