In Vanta, many tests require you to upload screenshots to satisfy unique requirements for your auditor. The Agentic Evidence Collection feature within the Vanta AI Agent streamlines the process by reading the evidence description for a document, determining what needs to be collected based on your Vanta environment and integrations, and then automatically navigating your apps to capture audit-ready screenshots on your behalf.
Preview: You can request access to this feature through your Customer Success Manager. We’re rolling this out gradually, so some documents may not support Agentic Evidence Collection just yet. Custom documents aren’t included, and for now the feature captures screenshot-based evidence only.
Collecting Screenshots
With the Agentic Evidence Collection feature, the Vanta AI Agent automatically collects audit-ready screenshot evidence for many document-based tests.
The agent reads the evidence description, determines where to collect screenshots based on your integrations, and navigates your apps to capture the right evidence. All you need to do is review the proposed plan, authenticate apps when prompted, and the AI Agent handles the rest, returning screenshots for you to confirm and upload.
To get collect a screenshot with the Vanta AI Agent:
From the Compliance section of your account, go to the Documents page
Open a document. Note: We’re rolling this out gradually, so Agentic Evidence Collection may or may not be available yet depending on the document.
In the Get started section, click Collect a screenshot. If you're renewing evidence, click New draft first to reveal this option.
In the chat panel that opens, select the prompt to kick off the process.
Wait while the Vanta AI Agent thinks out loud and generates a to-do list—it’s reviewing your integrations and planning how to capture the required evidence.
If needed, pause the agent to update instructions or ask it to revise the task list before proceeding.
When prompted, log in to the relevant account. For security purposes, we recommend using a dedicated, read-only service user account created specifically for the agent.
Confirm in the chat once you’ve logged in so the agent can continue. The agent will navigate through the app, take screenshots, and evaluate whether each image meets the document’s control requirements.
Review the collected screenshots and select which ones to add to the document.
Repeat the process for any additional accounts needed to complete the task list.
Refreshing or reloading the page will also end the chat session permanently. Even if you click Collect a screenshot afterward, the previous chat cannot be recovered and a new session will start instead.
Securing Login Credentials
For the safest experience, we recommend creating a dedicated, read-only service user account for the Vanta AI Agent to use during evidence collection. This helps keep your primary credentials secure and ensures the agent only has the minimal access needed to navigate your app and capture screenshots.
Once the service account is set up, you can provide its login credentials when prompted and continue using it for future evidence collections as long as its session cookie stays active.
You don’t need to log in again if your session cookie is still active—the Vanta AI Agent will continue using your existing session until it expires.
Re-opening Chat Windows
If you close the Vanta AI Agent chat during evidence collection, you can reopen it by clicking Collect a screenshot again within the same document. Chat sessions are document-specific, so you won’t see the same chat session if you click that button from a different document or when using the generic Vanta Agent button at the top of the page.
Refreshing or reloading the page will also end the chat session permanently. Even if you click Collect a screenshot afterward, the previous chat cannot be recovered and a new session will start instead.