AI Trust and Transparency Card

The AI Trust and Transparency Card lists who provides an agent, which models it runs on, who hosts it, which data it can see, and how that data is protected.

Open the card

Open an agent in Assist, in the standalone web app or embedded inside a Trimble product. Under the message box, select Transparency Card. To close the card, select X, press Esc, or click outside it. Opening the card does not change your conversation or the message you are typing.

The card describes the agent that is selected. After you switch agents, open the card again to see the new agent's details. The top of the card shows the agent's icon, name, and description, as written by the agent's creator. Select Show more to read a long description in full.

Below that, the card has three sections: About this AI, Security & Safeguards, and Data Handling & Accountability. Select a section heading to expand or collapse it. Hover over or focus the info icon next to a row to see what that row means.


The AI Trust and Transparency Card for an agent: its name and description, the About this AI rows, and the collapsed Security and Safeguards and Data Handling and Accountability sections

The agent's name, description, base models, and data sources come from how its creator set it up in Trimble Agent Studio.

About this AI

RowWhat it means
ProviderThe company that provides the agent
Use is optionalWhether you can do your work without using the agent
Model typeThe kind of AI model behind the agent, for example a generative model that writes new content from your prompts
Base model (or Base models)The models the agent can run on, the same list as the model picker in the message box. Select Show more when the list does not fit on one row
Hosted byWho runs the infrastructure the agent's models run on
Data sourcesEverything the agent can see while it works, besides the current conversation

Data sources

The Data sources row lists every source the agent can draw on besides the current conversation. Connectors and Web search are highlighted, because they can bring in content from outside Trimble. No additional or external data means the agent has none of these sources.

ChipWhat it means
Knowledge libraryThe agent searches documents in knowledge libraries. Select the chip to list the libraries by name.
Connector (or Connectors (N))The agent reaches other systems through connectors, which can include services outside Trimble. Select the chip to list them by name.
Web searchThe agent can search the public web
Past conversationsThe agent can read your own earlier conversations with it
Personal memoryThe agent can read your Trimble profile, such as your name, contact details, account, and organization, to personalize its answers

Security and safeguards


The AI Trust and Transparency Card with the Security and Safeguards and Data Handling and Accountability sections expanded
RowWhat it means
Useable for High-Risk ActivitiesWhether the agent is built for high-risk uses, such as hiring decisions, safety-critical systems, or biometric identification
Encryption at restWhether your data is encrypted whenever it is stored, even when the agent is not actively using it
Encryption in transitWhether your data is encrypted while it moves between you, the platform, and any connected tools
Model safeguardsWhether content-safety checks run on everything the agent outputs, to catch harmful or inappropriate responses before you see them
Human oversightWhether a person can review and step in before the agent's output leads to any real-world action

Data handling and accountability

RowWhat it means
Used to train modelWhether your data is used to train or fine-tune an AI model
Actions are loggedWhether everything the agent does is recorded, so its activity can be reviewed later if needed
System errorsWhether technical failures, like a timeout or a connector going offline, are tracked so they can be reviewed and fixed

For agent creators

You do not fill in the card separately. Assist builds it from your agent's configuration in Trimble Agent Studio.

  • Write an agent description that says what the agent is for and what it is not for. It is the first thing people read on the card.
  • Give knowledge libraries and connectors names people will recognize, because the card lists them by name.
  • Add only the sources the agent needs. Every library and connector you add appears on the card, as do the web search, past conversations, and My Profile built-in tools. Connectors and web search are flagged as external.

The card shows what the platform and you have put in place. It does not replace your own responsibilities as a creator.