Explicit inputs
The agent collects structured fields instead of sending an uncontrolled conversation blob.
Integrations
Connect conversations to approved business actions
Give the agent narrow, explicit tools instead of broad access. Start with read checks and structured outputs, then add confirmed write actions where they are justified.
6 detailed topics
English content
Integrations
An integration should make the next step clearer
Telp tools have an explicit description, input contract and permission policy. Credentials stay on the server, while operators can see what was attempted and where a workflow needs attention.
Controlled actions
Read checks, drafts and write actions can have different confirmation requirements.
Visible failures
A failed integration remains visible for review instead of silently pretending the action worked.
Explore details
All Integrations topics
Automation
Zapier, Make, n8n and automation platforms
Send structured conversation outcomes to Zapier, Make, n8n, Pipedream, Airtable or similar workflow tools.
Explore details
Customer data
CRM and help desk systems
Prepare or update customer records in systems such as HubSpot, Pipedrive, Salesforce, Zendesk or Intercom through approved connections.
Explore details
Commerce
WooCommerce and Wix commerce
Use native catalog and order capabilities for supported WooCommerce and Wix stores while keeping business rules in Telp.
Explore details
Back office
Google Workspace and email workflows
Use Google Sheets, Gmail or controlled email delivery as a practical operational output for smaller teams.
Explore detailsDeveloper
Webhooks and controlled REST tools
Connect internal systems through explicitly configured HTTP endpoints and structured input contracts.
Explore detailsAdvanced
MCP and custom server tools
Expose specialized operations when the business logic and authorization should remain owned by your server.
Explore detailsIntegrations
Choose the simplest reliable connection
Automation platforms are useful for a fast pilot. Native or direct server integrations fit flows that need stronger identity, authorization, product context or lifecycle control.
- What data the agent needs to read
- Which actions change business state
- Who must confirm sensitive writes
- How failures reach the responsible team
FAQ
Questions before you choose
How much access does Integrations receive?
Only the specific operations and fields exposed through approved tools. A connection does not give the agent unrestricted access to the underlying system.
Can write actions require a person to confirm them?
Yes. Sensitive operations can remain structured drafts until the platform receives explicit confirmation.
What happens when an integration fails?
The action is not presented as successful. Technical detail remains available in operational logs or debug surfaces, while the customer receives an appropriate next step.
Not sure where to start?
Share the channels you use and the customer work that repeats most often. We will suggest one focused first workflow.