Gambit

Gambit privacy policy

Effective 6 November 2026. This policy covers Gambit, Gambit Mail Sync, the Gmail add-on, the Outlook add-in, which works only with Gambit Beta, the website at gambit.humanr.ai and its sign-in and add-in pages at rainbook.humanr.ai, all provided by Human Renaissance Corporation.

Anyone Jira lets see a record sees all of it. Gambit's AI sees what you can see, and nothing more. Your data stays in your Atlassian cloud. Your records are Jira work items, and Gambit's indexes, history and settings are kept in Forge storage for your site.

1. Who is responsible for what

Your organisation decides what goes into Gambit and is the controller of that data. Human Renaissance Corporation provides the apps and acts only on your organisation's instructions, under our data processing agreement. Atlassian hosts the apps on Forge. Email from people outside your organisation is captured because your organisation chose to capture it; your organisation answers their requests, and Gambit gives your admins the tools to do it.

Human Renaissance Corporation runs no server that stores your CRM data or your email. What we process ourselves is limited to:

2. What the apps store

2.1 Jira work items in your site

Deals, leads, follow-ups, renewals, accounts, contacts, conversations, meetings and approvals, with their fields, comments, attachments and properties. Jira's permissions, backup and export apply to them, and they stay in Jira if you uninstall.

2.2 Forge storage for your site

Forge SQL, key-value storage and Object Store, hosted by Atlassian for your site only:

2.3 Email and calendar capture

When your organisation connects mailboxes, Gambit Mail Sync captures onto Jira work items: email threads someone on your team starts with, or replies to, a person outside your organisation, including people who are not in Gambit yet; email with known customers and open Leads matched by email only, never by company domain; meetings with known customers or open Leads; and internal threads a person chooses to Log from the Gmail sidebar (or, in Gambit Beta, the Outlook sidebar) after choosing a customer record. Team-organised meetings with outside guests are also captured before the guests are in Gambit: meetings the mailbox owner organised, or a colleague organised and the owner accepted or tentatively accepted. An outsider's invite with no known party is not captured just because the owner accepted it. Meetings with up to 150 attendees are captured; 151 or more are skipped.

"Someone on the team wrote" means their own sent copy; forwarding mail to your own personal address does not count. When someone on the team replies, up to 6 earlier messages from the 30 days before come with it. Inbound email from someone new that nobody on the team answers is not captured unless an admin turns on Capture all external email, which is off by default. History brought in when a mailbox is first connected (90 days by default) includes only mail with known customers.

Mail with more than 100 outside recipients is skipped unless a known party (a Contact, an Account's domain or an open Lead) is on the thread, or the thread is already being followed. Exactly 100 outside recipients is within the limit. Mailing-list headers still mark mail as bulk unless the sender is an exact Contact or open Lead.

Each person can keep their own correspondence out with a personal never-log list or a "Don't log" label, and admins can block addresses and domains. For each captured message and meeting, Gambit Mail Sync stores:

The people who write to your team are data subjects of your organisation, which is the controller. Captured email follows the record it is linked to: anyone who can open the account or deal in Jira can read its conversations. Conversations linked to no record are normally visible to your Gambit users. On sites with Legal-pack walls, new threads stay Private to the mailbox owner until linked; shared mailboxes there keep the known-customer rule. Any Gambit user who can open a thread can mark it Private, so that only its owner and the people Jira allows can open it. By default, that means the owner and members of rainbook-admins. On Jira Free everyone on your Jira site works in Gambit together and sees every record, including captured email; private deals and Private threads need Jira Standard or above.

Captured files are attachments, so Rovo and AI clients connected to Atlassian's MCP server find them for the people Jira lets see them. The mailbox owner, the thread's assignee or an admin can remove a captured message from Gambit; it can be restored for 30 days and is not captured again. Members of your rainbook-admins group can erase a person by email address, including people who never used Gambit.

2.4 Sensitive data

Gambit isn't built for health records, government ID numbers or other sensitive personal data, and the end user terms don't allow them. Payment card numbers and US Social Security numbers are masked wherever Gambit writes text: captured and logged email, calendar events, imports, Gambit's forms and AI output. Each number is replaced in place with a marker such as "[card ending 4242 removed]", and the rest of the text is kept. Attached files, file names, images and text typed directly in Jira's own screens are not scanned.

2.5 What the apps never collect

3. Gambit Mail Sync, the Gmail add-on and the Outlook add-in

Gambit Mail Sync is a separate app that alone talks to Google, and Gambit Mail Sync Beta also to Microsoft. Gambit itself connects to neither.

3.1 Google user data

Gambit's use and transfer to any other app of information received from Google APIs will adhere to Google API Services User Data Policy, including the Limited Use requirements.

Gambit uses Gmail and Google Calendar data only to show and capture your organisation's customer email and meetings in your Jira site and, on the Standard and Advanced editions, with Forge LLM (models hosted by Atlassian): to classify and summarise captured email threads, to summarise records and suggest next steps from them, and, when you press Draft reply in Gambit CRM for Gmail, to draft a reply to the open email, which is inserted only when you press Insert. It does not use this data for advertising, does not sell it, and does not let people read it except with your organisation's permission, for security purposes, to comply with law, or when the data is aggregated and anonymised for internal operations.

Gambit does not use data from Google Workspace APIs to develop, improve or train generalised AI or machine-learning models.

3.2 Microsoft data (Gambit Beta)

In Gambit Beta and Gambit Mail Sync Beta, Gambit uses Microsoft 365 mail and calendar data with Forge LLM (models hosted by Atlassian) to classify and summarise captured email threads, to summarise records and suggest next steps from them, and, when you press Draft reply in Gambit CRM for Outlook, to draft a reply to the open email, which is inserted only when you press Insert. Gambit does not use this data to develop, improve or train generalised AI or machine-learning models. Microsoft's own handling of your data is described in the Microsoft privacy statement.

4. AI

The AI Usage Policy lists what each feature sends.

5. Third parties, analytics and error reports

We don't sell personal data or share it for advertising. The services that process data for the apps are on the subprocessor list: Atlassian, which hosts the apps; Google (and, for Gambit Mail Sync Beta, Microsoft), for the mailboxes your organisation connects; Cloudflare, which hosts the website; Loops, for release and onboarding email to the people who sign up or hold a licence; PostHog, for product usage analytics; Sentry, for error reports; logo.dev, which receives a company's web domain to find its logo; and, from January 2027 and only when your admin turns each on, Salesforce, HubSpot and Pipedrive for imports, Apollo and People Data Labs for enrichment on your own key, the European Central Bank for exchange rates and SAM.gov for US federal contract opportunities. In a later release, and only when your admin connects your organisation's own account or key, Slack and Microsoft Teams receive the alerts you choose, Zapier and Make the events you choose for your automations, Brave Search the search terms for web research, and Anthropic, OpenAI or Microsoft Azure AI (Azure OpenAI or Azure AI Foundry) the AI requests you choose to run on your own key; nothing is sent to any of them until then.

Usage analytics (PostHog) record screens, features and timings, keyed by pseudonymous site and user IDs. Error reports (Sentry) record the error type, an error message stripped of names, addresses and content, and where in the code it happened, keyed by a scrambled site ID and never by person, and an import's reports also carry the import run ID and scrambled row references, never record content. Neither carries names, email addresses, record content or prompts. They are on when the apps are installed, and admins switch them off in Atlassian's admin controls for apps; the apps keep working without them.

6. The website

The website at gambit.humanr.ai counts page views with cookieless Cloudflare Web Analytics. Its forms send what you type to Human Renaissance's support desk or to our email lists, which Loops runs. The mailing-list form on gambit.humanr.ai stores your email address, the time you joined, the page you joined from, the page that sent you there, a one-way hash of your IP address and your browser's user agent in a Cloudflare database that only Human Renaissance Corporation can read. We use it only to email you about Gambit: the day it is on the Atlassian Marketplace, then occasional product news. Reply to any of those emails or write to privacy@gambit.humanr.ai to be removed, and we delete your row. A sign-up on the launch list also records which launch video and platform sent you and the time you agreed to email; Loops stores your email address, that source and the time, and your name and company if you give them. Cloudflare may set the __cf_bm and cf_clearance cookies to protect the website from automated traffic; they last 30 minutes and are not used for analytics or advertising. Apart from that list, the websites store no personal data.

7. Storage, isolation and residency

8. Retention and uninstall

Data Retention gives the full schedule.

9. Your rights

9.1 Access (GDPR Article 15)

Admins export everything Gambit holds about a person, found by email address, as a ZIP built in their own browser: the person's records, the email they sent, their files and a description of the contents.

9.2 Erasure (GDPR Article 17)

Admins erase a person by email address, including people who never used Gambit. One screen shows what will change; by default Gambit removes only that person's messages from shared threads and replaces their name and address with "[erased]" elsewhere. Copies in the Gambit Archive are rewritten without them within 30 days, and the address can be blocked from future capture. Atlassian users who leave are handled automatically: Gambit reports the account IDs it stores to Atlassian each week and clears closed accounts.

9.3 Data processing agreement

Our data processing agreement sets out how Human Renaissance Corporation processes personal data for your organisation. Atlassian processes data for the apps under Atlassian's data processing addendum.

9.4 Lawful basis

Your organisation decides the lawful basis for its CRM, usually its legitimate interest in managing customer relationships. Gambit only shows each person the data Jira lets them see. Human Renaissance Corporation processes its own customer and support data to provide the service and on its legitimate interest in supporting and improving the apps. Human Renaissance Corporation emails people on the Gambit mailing list on their consent, given on the sign-up form; they can withdraw it at any time.

10. California (CCPA)

11. Security

The trust centre describes the controls in full.

12. Changes to this policy

When this policy changes, we publish the new version on this page and change the effective date above. We email material changes to your licence's contacts before they apply.

13. Contact

14. Governing law

This policy is governed by the laws of the State of Delaware, United States, and its courts decide any dispute about it, without regard to conflict-of-law rules.