Instant DSAR scoping in Google Vault from Jira

PromptCreate

When a new issue is created in our Jira Privacy or DSAR intake project, run an agent that spins up the Google Vault side of the investigation and posts a scoping estimate back onto the ticket.

The trigger is a Jira issue-created webhook scoped to the intake project (Atlassian supports outgoing webhooks on issue create events, filtered by JQL). When it fires, take the new issue key from the webhook payload and use the Jira Get Issue operation to pull the full summary and description if you need more than the webhook body. Extract three things: the data subject's email, the request type (access, deletion, or portability), and any date range or corpus scope the requester specified. If the email is not present or is ambiguous, do not create a matter. Instead, use Jira Add Comment to ask the reporter to add the data subject email in a structured field, and stop.

Once you have the data subject email, call Google Vault Create Matter with a name that includes the Jira ticket key (for example "DSAR-482 — Jane Doe access request") and a short description linking back to the Jira issue. Keep the returned matter ID. Then for each in-scope corpus (default set: MAIL, DRIVE, HANGOUTS_CHAT, CALENDAR), call Google Vault Count Accounts against that matter with a query using method ACCOUNT and accountInfo.emails set to the data subject's email. Apply the requester's date range via startTime and endTime when they specified one, using RFC 3339 timestamps and the workspace time zone. Count Accounts returns a long-running Operation, so poll Google Vault Get Operation until done is true, then read the account count statistics off the response.

Finally, use Jira Add Comment on the original issue to post a scoping summary in Atlassian Document Format. Include: the request type and data subject you interpreted from the ticket, the date range you applied (or a note that you used the default), estimated message counts per corpus in a small table, a link to the Vault matter, and a suggested next step. Recommend 'proceed to export' when every corpus is under a reasonable threshold (start with 50,000 messages per corpus), and recommend 'narrow the query' when any corpus is above it, with a hint about which one is largest. Keep the tone factual and do not promise any legal outcome, since a human still owns the final decision. Do not run any export or delete operations — this workflow only scopes.

What does this prompt do?

Watches your Privacy or DSAR intake project in Jira and fires the moment a new request is filed
Reads the requester's email, request type, and any date range or scope hints out of the issue summary and description
Opens a dedicated Google Vault matter named after the ticket, then counts messages across Mail, Drive, Chat, and Calendar for the data subject
Comments a scoping summary back on the Jira issue with per-corpus counts, a link to the new Vault matter, and a suggested next step

What do I need to use this?

A Google Workspace account with Google Vault turned on
A Jira project where privacy or DSAR requests are filed
Ticket templates or a description convention that names the data subject and the type of request

How can I customize it?

Point it at a different Jira project or a saved filter for privacy tickets
Change which Google Vault corpora get scoped (add Groups, Voice, or Sites, or drop ones you don't use)
Adjust the default date range or the message-count threshold that flips the recommendation from 'proceed to export' to 'narrow the query'

FAQs

What kinds of privacy requests can this handle?

It works for any DSAR-style request that names a person: access, deletion, and portability requests under GDPR, CCPA, and similar regimes. The agent extracts the request type from the ticket and includes it in the scoping summary, but it does not fulfill the request. A human still reviews and decides what to export or delete.

Do I need a specific Google Workspace plan?

Google Vault is only available on Business Plus, Enterprise, and Education Standard/Plus editions. If your workspace has Vault turned on and the person running this workflow has Vault privileges, you're good to go.

What if the ticket doesn't include the data subject's email?

The agent tries to parse the email out of the summary and description. If it can't find one, it stops before creating a matter and comments back on the ticket asking the reporter to add the email so the scoping can run cleanly.

Does this actually export the data?

No. It only creates the matter and runs the account count so your privacy team gets an instant estimate. The scoping comment recommends a next step, but exporting still happens under human review in Vault, which keeps the audit trail clean.

How long does the scoping take?

Google Vault runs the account counts as background jobs, so the workflow waits for each corpus count to finish before commenting. For a typical account, the whole thing lands on the ticket within a few minutes, well ahead of the GDPR and CCPA response clocks.

Related templates

One triage console for every Jira service desk queue

Merge your IT, HR and Facilities queues into one list ranked by SLA time left, then reply, change status and escalate without ever opening Jira.

App

Work your whole Terraform approval queue from one board

Every Terraform run that needs a human, across every workspace, on one board with approve, discard, and plain English plan summaries.

App

Google Vault matter portfolio board with case brief intake

See every legal matter, hold, and custodian on one board, run the full matter lifecycle from it, and open new matters straight from a pasted case brief.

App

Google Vault query builder and export scoping workbench

Build a Vault search, preview how much data it would actually pull, tighten it until the scope is defensible, and only then run the export.

App

Google Vault custodian search and legal hold release console

Search any employee to see every legal matter and hold they sit on, check departing staff before their account is deleted, and release custodians with a logged reason.