What Are Agents ?
Agents in BonData are automated workflows that let you take action on data across your connected systems. Each Agent defines:
Conditions – when something should happen (e.g., "Accounts missing owner")
Actions – what should happen (e.g., "Send Slack notification", "Update record in Salesforce")
BonData constantly monitors your systems and executes your Agents when your defined rules are met—no code required.
Why Use Agents?
Automate Repetitive Tasks – Save time on manual exports, updates, and alerts.
Unify Disconnected Systems – Sync records, trigger actions, and gain visibility across CRMs, support tools, billing platforms, and more.
Stay in Control – Choose whether tasks run automatically or require approval before executing.
Types of Agents
Automation Agent– Triggers actions based on conditions like missing values or risk signals.
Migration Agent– Moves data between systems without storing it in BonData, with explicit field mapping.
Quick Start: Building Your First Agent
1. Access the Agent Builder
Go to the BonData dashboard → Agents → + Create New
2. Name & Describe
Give it a clear name (e.g., “Accounts Missing Owner”) and optional description.
3. First Node
To get started with your first node, click on the + Button
4. Get Data
A new blank Agent's first node will always be the Get Data node
5. Choose: Select Entity and Fields
Select the Entity and the Fields you want your Agent to monitor.
6. Define Conditions (Filters)
Add a filter from the + → Flow Basics
7. Add a Simple Filter
In the next window, select Simple Filter
8. Choose: Select Entity, Fields, Operator and Condition
Select the Entity, Fields, Operator and Condition you want to filter results by
AND Logic
To Filter using the AND Logic, simply add another Simple Filter to your next node from the Then + option
a.
b.
OR Logic
To Filter using the OR Logic, add your condition to the existing node from the + Add condition option
a.
b.
9. Set Actions
Decide what to do when filters are matched:
Field Mapping
Choose Execution Mode
Automatic: Runs tasks (actions) immediately
Manual: Requires your approval for each task
Preview Your Agents (Dry Run)
Before running it live, click the "Test" button (top right, next to Run and Save). This creates a preview of the tasks without performing actions—ideal for reviewing what data would be created or updated. Tasks will be shown in the panel on the right hand side of the screen.
Make It Live
To turn your Agents into a running, automated workflow:Click on Active (top right)
Choose Run once or Schedule to run periodically
"Run once" will run the agent once, creating tasks. Depending on how the tasks were set, they will either run automatically once or wait for you to run the tasks.
"Schedule to run periodically" allows you to define an automated frequency of running the task (e.g., every 15 mins or custom cron expression)
Choose whether to Allow task duplication or not
By default disabled.
Click Apply to activate the Agent for the first time
Monitoring & Managing Agents
See Task Previews: Right menu on the Agent page shows upcoming tasks
Track Task Execution:
More icon -> Agent Dashbord next to Activate
Monitor task statuses: Running, Pending, Failed, Success
Analyze Matrix results by clicking on View Results for Get Data nodes, Bonds nodes, Filter nodes...
Click on a task from the task list to review task outputs
Mentions
Use "Mentions" to dynamically insert field values from one system into another (e.g., take “Region” from a CSV and insert it into Salesforce's BillingState).
You can also use:
Conditional Mentions – Apply logic (e.g., if 1 → “{variable from Region field}”, else → Static value: "Other")
Default Values – Use fallback (default) values when source fields are blank
Use Cases
Data Cleanup – Flag or export incomplete records
Data Sync – Keep status fields aligned across tools
Alerts – Notify teams when rules are met, or periodically with updated information
Migrations – Move records from a legacy system into a new platform with field mapping
Best Practices
Start small: test with a simple condition and action.
Use Test (Dry Run) to validate before activating.
Periodically audit filters and field mappings.
Monitor task volumes to catch overly broad rules.
FAQs
Can I add multiple actions?
Yes. One Agent can perform several actions per matched record.
Can I test without making changes?
Yes. Use Test or set execution to Manual.
Can I work across systems?
Absolutely. Agents can reference multiple integrations.
Can I undo a mistake?
Check the Task Logs for affected records. Since BonData doesn’t store full datasets, rollback is manual in the original systems.




























