Tajo 101: the agent team behind a sync
In short: one agent builds the sync, one runs it, and seven work on the contacts and orders it lands. Nothing personal reaches Brevo without a consent rule, and no sync goes live or agent send goes out without your approval.
Tajo builds integrations between the systems you already use. Every guide here starts with a sync: records from a source system mapped into a destination. Brevo is the first destination for email and SMS. HubSpot contacts is the second.
The team
Every Tajo workspace starts with the same set of agents. These nine are the ones sync work runs through. One card per agent, with what each never does: Tajo agents.
| Agent | Its job | What it works with | What it never does on its own |
|---|---|---|---|
| Brevo Integration Architect | Turns a source into a sync rule: samples real records, proposes the field map, validates it, dry-runs it, saves a draft | The source connector, Brevo’s endpoint catalogue, sync-rule drafts | Activate a sync. It asks for your approval and activates only with the approval it was given |
| Sync Operator | Runs active sync rules and reports what was pulled, loaded and refused | Your active sync rules | Change a rule |
| Segment Builder | Turns a description of who you want into a contact list | Synced contacts, per-channel consent | Send anything |
| Compliance Reviewer | Checks a draft before it goes out: consent, unsubscribe, sender domain, content red flags | The draft and the channel | Approve a send for you |
| Engagement Designer | Drafts email, SMS and WhatsApp messages and templates | Templates and campaign drafts | Send. Drafts wait for review |
| Performance Analyst | Reads open, click and unsubscribe rates, flow completion and segment health over up to 90 days, and writes recommendations | Your workspace’s engagement results | Act on its own recommendations |
| Orchestrator | Takes a free-form request and chains the other tools to finish it | Everything above | Skip a gate. It and Win-back are the only two that can schedule a send, and every send stops for your approval |
| Win-back | Finds customers who ordered before but not lately, sizes that audience and drafts a win-back campaign for it | Order events synced from your store, per-channel consent | Send without your approval, or create a discount code. It has no tool for one |
| Customer Intelligence | Scores customers by recency, frequency and order value and reports who is buying, slowing down or gone quiet | Order events synced from your store | Change a contact, a segment or a connected app. Every tool it holds reads |
How a sync gets built
Ask the Brevo Integration Architect for a sync and it works through the same six steps every time.
- Sample. It pulls a few real records from the source, so the field map is built from data you have.
- Map. It proposes which source field goes to which Brevo field.
- Validate and dry-run. It checks every mapped field against the Brevo action the rule writes to, then dry-runs the rule without writing anything.
- Save a draft. The rule exists, but nothing moves.
- Ask. It requests your approval to publish. The request expires after 30 minutes.
- Activate. Only with that approval.
Ready-made rules exist for common sources. You pick one under Connections → Syncs and mappings, and it becomes a draft you can edit before publishing.
What stays with you
Four gates hold whichever agent is working.
- Publishing a sync. Every sync rule starts as a draft. Going live takes your approval.
- Consent before the data moves. A sync that writes contacts or events to Brevo cannot be published unless it filters on consent and suppression. At run time a record is refused when its source cannot prove consent. Missing consent means no write.
- Consent before a message goes out. A marketing send needs an explicit opt-in on the channel it is sent on. An email opt-in never covers SMS. Unknown means no send.
- Sends. A send an agent schedules always waits for your approval. Any send to 50 or more contacts does too. An approval covers the exact action you saw and nothing else.
Pending decisions collect under Approvals.
Routines
Anything you ask on the home screen with a schedule in it becomes a routine that runs the same agent again. The schedule words map to fixed times, all in UTC:
| You write | It runs |
|---|---|
| ”every hour” | At the top of each hour |
| ”every weekday” | 09:00, Monday to Friday |
| ”every Monday” (or any weekday) | 09:00 on that day |
| ”every week” | 09:00 on Mondays |
| ”every month” | 09:00 on the 1st |
| anything else, such as “every morning” | 09:00 daily |
The sync itself does not wait for a routine. Tajo checks for due contact syncs every minute and runs each active rule at most every 15 minutes.
Caution
Coming: routines started by an event, such as a new order, are not built yet. Today every routine runs on a schedule.
Prompts that work today
- To the Brevo Integration Architect: “Build a sync from Shopify customers to Brevo contacts. Show me the field map and a dry run before you save the draft.”
- To the Sync Operator: “Run the Shopify customers sync now and tell me how many records loaded and how many failed.”
- To the Segment Builder: “Make a list of contacts who opted in to email.”
- To the Compliance Reviewer: “Check this draft for an unsubscribe link and consent problems before I approve it.”
- To the Performance Analyst: “Look at the last 30 days of email results and tell me what to change.”
- On the home screen: “Every Monday, review last week’s email results and tell me what to change.”
Start
Tajo workspaces are not open for self-serve sign-up yet. Sign up for early access with your first name and work email: you confirm by email, then we send access details when a workspace can be opened for you. If you run a Shopify store, the next guide is Tajo for Shopify stores.