Admin Use Cases
Practical admin workflows for object setup, imports, exports, workspace configuration, requirements, agents, and inbox operations.
This guide explains the most important administrator workflows in Recerc, with practical steps, safeguards, and recommended operating patterns.
Quick Links
- Workspace Settings
- Object Definitions
- Object Import
- Exports
- Global Agent Prompt
- Custom Agents
- Inbox
- Integrations
- Webhooks Docs
1) Configure Objects
Use object settings to define structure, validation behavior, and navigation for both default and custom objects.
- Go to Object Definitions.
- Open a specific object definition at
/w/dashboard/settings/objects/<object-handle>. - Example: Configure Products Object
Where field types are configured
Field types are defined inside each object configuration screen (Add/Edit field). Available field types include:
textnumbercheckboxdatetimestampselectmultiSelectlocationcustom_object_relationobject(nested fields)
Default objects vs custom objects
- Default objects (for example
products,customers, and other system objects) can be configured to fit your workspace while preserving core platform behavior. - Custom objects let you model workspace-specific entities and relationships that do not exist out of the box.
Recommendation Keep handles stable after rollout. Changing naming conventions frequently can break downstream import mappings and automation assumptions.
2) Bulk Import Objects
Use object import to append or upsert records at scale.
- Go to:
/w/dashboard/objects/<object-handle>/import
- Upload CSV/XLSX.
- Review mapping.
- Preview transformed rows.
- Run import.
About the template
You can download a template directly in the wizard. It is useful, but your file does not need to be an exact template clone. Auto-map can infer mappings from column names and sample rows.
Example minimal CSV:
sku,title,basePrice.amount,basePrice.currencyCode
SKU-001,Industrial Drill,129.9,USD
SKU-002,Safety Helmet,19.5,USDUpsert key (reference key)
- When Upsert is enabled, Recerc tries to match incoming rows by the selected Reference Key.
- Typical pattern:
- Products:
skuis the common reference key. - Custom objects:
externalIdis the safest long-term key.
- Products:
- If a match is found, the record is updated.
- If no match is found, a new record is created.
Auto-map and additional instructions
In the mapping step:
- Use Auto map to generate initial mappings automatically.
- Use the options button next to Auto map to add additional instructions (for example: "prioritize SKU matching, avoid expression fields unless required").
Data behavior during import (important)
- Records not present in the uploaded file are not deleted.
- Fields that are not mapped / not provided remain unchanged.
- Fields that are present in your file with a blank value can overwrite existing values as blank.
Warning Blank cells in mapped columns are treated as incoming values. If you want to preserve existing values, do not map that column or remove it from the file.
3) Export Objects
Exports are asynchronous jobs and require elevated permissions.
- Go to Exports
- Create a new export request.
- Wait for job completion.
- Download from export history.
Access and filters
- You need export permissions (workspace management + admin role) to access and create exports.
- You can define:
- Maximum rows (
limit) - Updated since timestamp (export only rows updated after a given date/time)
- Format (
csvorxlsx)
- Maximum rows (
Recommendation For large datasets, always set both a row limit and an updated-since window to reduce job time and improve operational repeatability.
4) Workspace Configuration
Use Workspace Settings for company and operational defaults.
Configurable fields
shortNamecompanyNamecompanyDescriptionwebsitephoneNumberaddresscontactEmailcontactFaxlogoUrldefaultPriceListIdorderNumberPrefixorderNumberBasecurrencyCodeinvoiceDefaultFooterTextdefaultPaymentTerms
Global agent prompt
Use Global Agent Prompt to define workspace-wide agent instructions.
- This prompt is injected as shared workspace instructions for agent interactions.
- It applies globally, while each custom agent can still add its own agent-specific prompt.
5) Requirements Configuration
Requirements let you enforce operational checks before sales progression.
- Configure in Workspace Settings.
- Enable/disable requirements globally.
- Create requirement templates with:
- key
- title
- description
- enabled flag
How requirements are used
- Enabled templates are added as pending requirements on new orders.
- Teams can approve/reject each requirement with an optional reason.
- In review flows, pending requirements can block progression until approved.
Recommendation Keep requirement templates short, measurable, and auditable (for example: "PO attached", "Margin approved", "Compliance check complete").
6) Custom Agents for Specific Tasks
Use Custom Agents to create specialized assistants for recurring workflows.
Each agent can have:
- Name and description
- Agent-specific prompt
- Optional custom model override
Common admin use cases:
- Converting complex quote formats into structured records
- Generating recurring commercial or operations reports
- Summarizing customer communications and extracting next actions
Recommendation Keep prompts explicit about output format and decision criteria. Short, structured prompts are easier to maintain and debug over time.
7) Inbox and Channel Integrations
Use Inbox as the operational center for incoming customer requests and agent-assisted responses.
Supported channel patterns
- Email inbox webhooks (incoming emails can create threads and trigger agent responses)
- WhatsApp Business integration via Integrations
- API-driven inbox ingestion (for example webhook routes such as
agent-inbox)
Operational model
- Incoming messages create/append inbox threads.
- Agents can be triggered automatically for configured channels and manually from the thread view.
- Teams can review and continue conversations from a unified thread model.
Warning Before enabling fully automated replies in production, validate tone, escalation rules, and fallback behavior for edge cases (ambiguous requests, missing data, policy-sensitive topics).