BrainChat · AI OS para empresas
BrainChat for large operations.
Processes, support and AI agents in the same environment, under your company's access rules.
Why companies choose it
- Simple to govern: The same permission applies in the menu, on screen and in the database. What a person can't see never reaches their browser.
- Easy to use: Each area works in its own view: support queue, process board or portal. No need to train everyone on everything.
- Clear to lead: Dashboards calculated over the whole base, with response time, volume and results by channel, team and stage.
Your company's way
- Layered access: Roles in the organization and in each channel, areas granted per person and item-by-item sharing.
- Process rules: Required fields per stage, allowed path and edit lock, also checked on the server.
- Agent with limits: The agent only runs the actions you configured and hands over to a person when the rule says so.
- Portal per person: Customers and partners sign in to the portal and each one sees only the records linked to them.
Enterprise-ready: customize
- Data model: Relational tables. Formula fields. Unified table from several sources.
- Processes: Stages and advance rules. Forms and record actions. Automation with Flow.
- Dashboards and portals: Studios on top of tables. Portal on your own domain. Home screen by audience.
Enterprise-ready: control
- Identity: Two-step verification. Passkey sign-in. Shared computer mode. Sign-in and sign-out log.
- Access: Permission per area and per action. Roles in the organization and channel. Sharing per item. View by department.
- Agent limits: Actions configured one by one. Handover to a person by rule. Test lab before publishing. AI usage per resource.
Enterprise-ready: compliance
- Data: Database and files in São Paulo. Encryption in transit and at rest. Daily backups. Export from your own account.
- LGPD: Data protection officer (DPO) appointed. Data subject channel with protocol and deadline. Public list of subprocessors. Acceptance with hash and timestamp.
- Operations: Public status page. Incident notice inside the platform. History of each automation. Continuous monitoring.
Enterprise-ready: connect
- Channels: WhatsApp through the official API. Email on your own domain. Live Chat on your website. Calls through WhatsApp.
- Systems: Read and write API. Conversation and data webhooks. Pipefy inside the card. HTTP calls in Flow.
- Open AI: MCP connector. SDK and embed. Customer's own AI key in the agent.
Open by design
- Screen, API and connector: What you do on screen you can also do through the API and the MCP connector.
- One access rule: The connector acts as the person who connected it and sees only what that person sees.
- Our agent or yours: Use BrainChat agents or plug in the AI assistant your company already adopted.
- Your data stays yours: Export any table, or everything, whenever you want.
Start with one process
- Design: We pick a real process and define tables, stages and who accesses what.
- Data: We import spreadsheets and connect source systems. A pilot team starts using it.
- Build: Agents, automations, dashboards and integrations tested with real cases.
- Operation: Teams join in stages and the next process starts on the same base.
Frequently asked questions
Can we bring what we already have?
Yes. Spreadsheets come in through import with column mapping, and systems come in through API or sync, like Pipefy. On import, a record that already exists is updated, never duplicated.
Who sees what?
Each person sees only the areas granted to them and the items they created or that were shared with them. The rule is checked in the database, not just on screen.
Where is the data?
The database and files are in São Paulo. The Trust Center lists the subprocessors, the country of each and the channel for data subject requests.
Does the AI learn from our data?
No. Customer data is not used to train our own models, and the AI receives only what it needs to answer.