The Regulated pack

The Regulated pack adds the controls a regulator, auditor or professional-conduct rule asks for on top of Jira's own permissions. It is built for healthcare and life sciences, financial services, government and defence, and legal teams: keep email content off Jira work items, prove that activity history has not been changed, stop sensitive personal data at the door, and keep confidential business terms in a space only named people can open.

The Regulated pack is in early access from 15 December 2026. It is optional and off by default: an admin turns it on by installing it. It installs from Settings › Packs on Rainbook's Advanced edition, on Jira Standard or above. To try it, start an Advanced trial and install it there, or write to support@rainbook.humanr.ai to join the first customers we work with each week.

Profiles, preview and removal

At install you pick one or more profiles: Healthcare and life sciences, Financial services, Government and defence, and Legal. Each profile sets the right detector actions and options for that kind of team, and an industry pack you already use suggests its profile for you.

With the pack, access is group-only: only members of the Rainbook users group use Rainbook, even on a site where everyone who can open the CRM space would otherwise use it. Only members of rainbook-admins administer Rainbook; Jira admins are not Rainbook admins here unless they belong to that group. Before anything changes, the install preview lists every space, group, security level and setting the pack will create, in plain sentences, with one Install button. After install, a setup checklist walks your admins through the recommended next steps, such as choosing whether Rovo can search Jira and narrowing who sees captured email; nothing on it blocks you. Admins change the pack's settings or remove it from the same Settings › Packs page.

The regulated store

With the regulated store on, which is the default, email bodies and files stay out of Jira work items:

• Captured email bodies and files go to Rainbook's regulated store in the Forge storage Atlassian hosts for your site, carried there only inside Rainbook's own signed messages.
• Each Conversation keeps a short note in Jira with From, To, Cc, the sent time and the number of attachments, so people can still see that an email happened. Subjects are never written to Jira.
• AI summaries, next steps and health reasons stay in Rainbook's storage; the numeric Health score stays on its Jira field.
• The inbox sidebars show the contact, account, owner and open deals, with a link to open the rest in Rainbook.
• New threads that someone on your team wrote stay Private to the mailbox owner until they are linked to a record, and a shared mailbox keeps capturing only mail with known customers.
• The Rainbook agent in Rovo is off by default, email summaries for outside AI clients are off, and the nightly Rainbook Archive is a download from the admin page rather than an attachment in Jira.

Download full regulated export on the pack's settings page builds a ZIP in your browser with every regulated email body and file and all of your sealed history, ready for your own archive.

Sealed history

From the moment the pack is installed, every activity-history entry is linked to the one before it in a hash chain, so any change to a past entry shows. Each day is sealed into a segment signed with your site's own Ed25519 key; the public key is shown on the activity history page and included in every export.

• Verify chain on the activity history page checks any date range and names the first entry that does not match.
• Export sealed days downloads any date range as a ZIP of segments, their manifests, the public key and a VERIFY.md with the steps below, so you can check it offline and keep it in your own write-once storage.

How to verify a sealed export

1. Take the entries in order of their chain number. The first entry's previous hash is the SHA-256 of frontier-audit-genesis: followed by your installation ID, a colon and 1.
2. For each entry, compute SHA-256 of its previous hash followed by the entry without its hash field, written as canonical JSON: UTF-8, keys sorted, no whitespace, times in ISO 8601 with milliseconds and Z. The result must equal the entry's hash, and the next entry's previous hash.
3. For each daily segment, check the manifest's entry count, first and last chain numbers, first previous hash, last hash and SHA-256 of the segment file, and check the manifest's Ed25519 signature with the public key in the export.
4. Each segment's first previous hash must equal the previous segment's last hash, so no day is missing.

Sensitive-data detectors

The pack turns on Rainbook's whole sensitive-data detector library, at capture and at import. A blocked message or file is not written; your admins get a notice that names the kind of data found, never the data itself.

Rainbook is built for customer relationships, not for special-category personal data, and the end user terms don't allow it; with the pack, the detectors keep it out. Without the pack, Rainbook still masks payment card and US Social Security numbers wherever it writes.

The Confidential space

The pack creates a restricted Confidential space in your Jira, with its own permission and security schemes, a rainbook-confidential group and a security level, for commercial terms your admins choose to restrict. Nothing in the Confidential space is ever sent to Rainbook's AI: not to Ask, summaries, drafts, agents or the Rainbook agent in Rovo.

Confidential records live in a restricted Confidential space in your Jira. People who can open the Confidential space can also see its contents in Rovo search and chat. Rainbook's AI never receives them. To keep them out of Rovo completely, block Rovo for Jira in Atlassian Administration.

Jira site admins can still change the space's schemes, as with any Jira space.

Ethical walls

With the Financial services and Legal profiles, a private deal can be switched to Ethical wall. Only the people it is shared with can open it and its email; Rainbook admins can't unless they are named on it. Rainbook's AI never uses its content on any other record, and the deal stays out of activity pings and shared feeds. Jira site admins can still change the space's schemes, so the wall is as strong as your Jira administration.

Rovo and AI clients

Rovo search and chat read Jira work items, comments and attachments under Jira permissions. Your admins can block Rovo for all of Jira, or remove Rovo from the groups that must not search CRM content, in Atlassian Administration → Rovo → Rovo access; Rovo can't be blocked for one space only. For AI clients connected to Atlassian's MCP server, we recommend an Atlassian data security policy that blocks app access to the Activity and Confidential spaces. Anything a person sends to an outside AI client, such as ChatGPT, is covered by your organisation's terms with that provider.

Data inventory

What the pack adds to the data Rainbook keeps. Everything else is in the data inventory on the trust centre.

Uninstalling

With the Regulated pack, email bodies and files live in Rainbook's regulated store, outside your Jira work items, and imported history older than 90 days lives in Rainbook's storage. Uninstalling removes both from your site; Atlassian can restore them to a reinstall you request within 21 days. Run Download full regulated export first; it includes your imported history. Confidential records stay in your restricted Confidential space.

If you uninstall Rainbook, email content and audit history held by Rainbook are deleted after Atlassian's recovery period. Download a full export first.

For EU financial entities

The DORA addendum sets out the contractual terms the Digital Operational Resilience Act asks of an ICT service provider, and the register-of-information data for your register.

Coming next

Each of these joins the pack in a minor version, and this page gains its section the day it does.

• Protected fields on the record page: confidential values shown on the deal or account page only to people who can open their record in the Confidential space, and an install step that moves the fields you choose there. Protect your sets before your team enters data: Rainbook hides a protected field's value from everyone without access, and from search, Ask, AI and exports, but Jira keeps its own change history for every work item, so a value entered before the field was protected stays in that work item's History tab, visible to anyone who can open the work item, and Rainbook cannot edit Jira's history.
• Retention schedules and legal hold: set how long captured email, meetings, sealed history, AI drafts and nightly archives are kept, and hold chosen records so nothing in scope is deleted.
• Adverse-event intake: with the Healthcare and life sciences profile, possible adverse events and product complaints held back from capture and routed to your safety team.
• Regulated email backfill: bring past email into the regulated store from January.

Questions

Write to privacy@rainbook.humanr.ai or security@rainbook.humanr.ai.