SpaceXAI opens shared Grok Team Bots in public beta
By AgentRiot Editorial
The new team layer puts common instructions and tools behind a Bot that colleagues can use separately or bring into Slack. Access is limited to Teams and Enterprise plans.

SpaceXAI has launched Team Bots, a way to share a Grok Bot with coworkers rather than have each person assemble the same role from scratch. The September 28 announcement says the feature is in public beta for Teams and Enterprise plans. Its practical pitch is a shared account, project or function specialist that can carry the team's reference material and work through connected applications.
A Team Bot combines files, instructions and skills as context; plugins for applications such as Salesforce, Notion and GitHub; credentials for third-party APIs without a plugin; and memories that accumulate as the Bot works. Plugins can be connected by individuals or configured for the team, SpaceXAI says. That distinction matters: a shared Bot is not a claim that every member automatically has the same access to every connected system.
The announcement describes two kinds of continuity at once. The team shares the Bot's role and expertise, while each person's conversations remain private and the Bot keeps separate context and memories for each user. Team knowledge can still improve across interactions: SpaceXAI says its Data Bot remembers corrections to queries so a lesson from one person can improve answers for others. The company does not explain the exact boundary between those shared improvements and private per-user memories in the launch post. Teams handling sensitive records should not mistake the short description for a complete data-governance specification.
One Bot in the channel, not one chat for everyone
Each Team Bot can have a Slack handle and be invited into a channel, where colleagues can ask questions, add context and see its replies, according to the announcement. That is a different collaboration surface from an individual's private conversation with the Bot. Teams should decide which requests belong in a visible channel and which belong in a private thread; the launch description of private conversations does not make channel replies private.
SpaceXAI describes an account Bot that reviews recent Gong calls, Notion documents, Slack threads and company news overnight, then posts a morning brief for sales and customer-success staff. Its engineering example follows project decisions in Slack, files bug tickets and delegates defined fixes to Cloud Agents. Those are company-reported internal workflows, not independent performance tests or a promise that every customer deployment will behave the same way.
The data example makes the access question concrete. SpaceXAI says its Data Bot uses shared, read-only credentials to query approved Databricks tables and connects to Datadog, Hex and Statsig. Read-only is the stated configuration of that example, not a universal limit on Team Bot credentials. The launch post calls credentials a means of securely accessing third-party APIs but does not specify how team-shared secrets are stored, scoped or rotated. Admins need those particulars before treating a Bot as a general-purpose path into sensitive systems.
A team layer on an existing Bot
This is not the first launch of Grok Bot. SpaceXAI introduced the underlying product on August 11 with persistent Bots, computer and app use, routines, and conversations across desktop and mobile. That announcement described individual-plan access and a narrower Teams beta, with Enterprise users directed to a waitlist. The September release adds a shared role, team context and Slack participation, and now says Team Bots are in public beta on Teams and Enterprise plans. The two announcements describe different cuts of the product; the older eligibility list should not be read as today's Team Bots entitlement.
Neither the Team Bots announcement nor its beta wording gives a separate Team Bots price or a guaranteed allowance. SpaceXAI points customers toward pre-built Bots for sales, product management, marketing and analytics, but the purchase and rollout decision turns on a narrower question: which team members and integrations will have access to a shared Bot, and what work should remain subject to human review?

