A team workflow, with real sign-off.
BAT gives a marketing team five brand roles, a real approval state machine, and an audit log on every review. Creators draft, managers approve, and nothing reaches a platform until the right person signs off. No post-publish recall to clean up afterward, because review happens before anything goes live.
Five roles, one capability matrix.
Access in BAT is role-based, not all-or-nothing. Every brand workspace has five roles, and the capability matrix decides who can draft, who can approve, and who can change settings. The hierarchy is enforced on every request, so a junior account cannot quietly grant itself more power.
Owner and Admin
Full control of the brand workspace: billing, integrations, member access, and final approvals. Owners and Admins set the guardrails everyone else works inside.
Manager
Approves or rejects drafts and runs the review queue, without touching billing or workspace-level settings. The reviewer role for a working marketing team.
Creator
Drafts posts and submits them for review. Creators move work into PENDING_REVIEW but cannot approve their own output, so a second set of eyes is structural, not a habit you hope holds.
Member
Chat-only by default: explore the brand brain, ask the agent for analysis, and propose ideas, without publish or approval rights. A safe role for stakeholders who need visibility, not control.
Approval is a real state machine, not a checkbox.
Every post draft moves through an explicit lifecycle. Each transition is a recorded event, not a flag someone can flip and forget, and rejection is a first-class step with notes and a clean path back to the queue.
- DRAFT. A Creator (or the AI agent, drafting from the brand brain) writes a post per platform in the brand's voice. It sits in DRAFT until someone submits it.
- PENDING_REVIEW. Submitting moves the draft into the review queue. It is now waiting on a Manager, Admin, or Owner. Creators cannot approve their own work, so the queue is a genuine gate.
- REJECTED, with a note. A reviewer can reject with a written reason. The note travels with the post, so the Creator sees exactly what to fix instead of guessing, then resubmits the same post back into PENDING_REVIEW.
- APPROVED, then SCHEDULED. Approved posts hand off to the real scheduled-publish worker, which atomically claims due posts so a retry or a race can never double-publish. Approval is the only door to the scheduler.
Review before publish, not damage control after.
Plenty of tools let you publish first and offer a fragile window to recall a post. BAT does not work that way, and it does not pretend to. There is no post-publish recall, because the review happens before anything ships. Once a post is live, it is live, the same as posting by hand.
That is the honest trade. It means the approval queue is the thing that protects you, so BAT makes it strict: the role hierarchy includes privilege-escalation guards, so an account can only grant roles below its own, and it cannot promote itself to gain approval rights. The gate cannot be walked around from inside.
For a marketing team, that is the right default. The cost of a wrong post on a brand account is a public mistake. Catching it in PENDING_REVIEW costs a rejection note and a resubmit.
Evidence everyone can trust.
A review queue only works if reviewers can trust what they are approving. BAT drafts from a persistent per-brand brain and attaches provenance to outputs, so a Manager is signing off on something traceable, not a confident guess.
Drafts from the brand brain
Every draft is generated from the brand's own DNA, audience personas, competitors, and content pillars, the same persistent knowledge base for everyone on the team. The voice is consistent because the source is shared.
Provenance on outputs
BAT labels outputs VERIFIED, PARTIAL, BLOCKED, or RAW so a reviewer can see whether a claim traces back to a real workspace source or needs a second look before it ships.
An audit log on every review
Every approval, rejection, and resubmit is logged, and destructive brain or Drive changes from chat require explicit approval before they persist. You can reconstruct who decided what, and when.
One queue, every platform.
Approved posts publish natively to all 8 networks from one calendar, each in its native format, with AI-suggested send times drawn from the brand's own engagement history. The team reviews once; BAT handles the per-platform shaping.
Where this fits a marketing team, and where it doesn't.
BAT is built for an in-house team running one brand or a small set of brands with real review discipline. It is worth being clear about the edges.
- It is a brand-side workflow, not an agency white-label suite. If your job is reselling reports under your own logo for many clients, our white-label and agency reporting are not as mature as a dedicated agency tool. See /for/agencies for the honest version of that picture.
- Pricing is flat per brand workspace, not per seat. Brands Starter is $100/mo for one brand and includes 3 members, so adding reviewers does not inflate the bill. Compare that to per-seat suites where every approver is another line item.
- Approval is the safeguard, so design your roles deliberately. Decide up front who reviews. Members are chat-only, Creators submit, and only Owner, Admin, and Manager can approve. The matrix is fine-grained on purpose.
- Analytics depth varies by platform. First-party analytics are strongest for LinkedIn, Instagram, and Facebook; X, TikTok, Threads, and Pinterest data is gathered via scrapers. See /social-media-analytics-tool for what each surface actually reports.
Frequently asked questions.
How does the social media approval workflow work?
Every post draft moves through an explicit state machine: DRAFT, then PENDING_REVIEW when a Creator submits it, then either APPROVED or REJECTED by a Manager, Admin, or Owner. Rejections carry a written note, and the Creator can fix and resubmit the same post back into PENDING_REVIEW. Approved posts hand off to the scheduler. Review always happens before publish.
What are the five team roles, and who can approve posts?
Owner, Admin, Manager, Creator, and Member, governed by a fine-grained capability matrix. Owners and Admins control billing, integrations, and access; Managers run the review queue; Creators draft and submit; Members are chat-only by default. Only Owner, Admin, and Manager can approve. A role hierarchy with privilege-escalation guards means an account can only grant roles below its own.
Can a teammate approve their own drafts?
No. Creators move drafts into PENDING_REVIEW but cannot approve them; approval requires a Manager, Admin, or Owner. That separation is structural, so a second set of eyes is the default, not a process you have to remember to follow.
Is there an audit trail for reviews?
Yes. Every review transition (approve, reject, resubmit) is logged, and destructive brain or Drive changes from chat require explicit approval before they persist. You can reconstruct who decided what and when.
Can we recall a post after it publishes?
No. BAT is approval-gated before publish, not after. There is no post-publish recall, because the review queue is the safeguard and it runs before anything goes live. Once a post is published it behaves like any normal post on that platform.
How much does it cost for a team, and is there a free plan?
There is no free tier. BAT runs on a 14-day, 200-credit trial with no card required, then starts at $100/mo for Brands Starter, which covers one brand and 3 members. Pricing is flat per brand workspace, not per seat, so adding reviewers does not raise the price. Frame it against the cost of a hire or an agency retainer, not against a free scheduler.
Give your team a real sign-off.
Five roles, a true approval state machine, evidence-linked drafts, and an audit log on every review. Start free for 14 days, no card required.