Retriever AI Client Workflow Deployment (Delivery)

Retriever AI Client Workflow Deployment (Delivery)

1. 01

   Log into the Retriever AI cloud platform and navigate to the Agents section.

   Ensure you are on the Team plan ($99.99/month, 10,000 credits) to have enough credits for multi-client operations.

2. 02

   Create a new agent by clicking 'New Agent' and name it after the client project.

   Use a naming convention like 'ClientName_WorkflowType' to keep agents organized across accounts.

3. 03

   In the agent configuration screen, write a natural language prompt describing the browser-based task.

   For example, 'Scrape all salon names, addresses, and phone numbers from Yelp for Austin, TX' or 'Enrich these 500 leads with LinkedIn URLs'.

4. 04

   Select the output format (structured data, CSV, JSON) and set up any required triggers or webhooks.

   Use the 'Triggers & webhooks' feature to automate delivery to the client's spreadsheet or CRM.

5. 05

   Test the agent using the interactive browser extension for real-time validation.

   Run a small sample first to confirm the agent handles authenticated sites like LinkedIn or Airbnb if needed.

6. 06

   Schedule the agent for recurring runs via the 'Schedules' tab.

   Set daily or weekly pulls depending on the client's monitoring or enrichment needs.

7. 07

   Monitor credit consumption in the usage dashboard to avoid exceeding the monthly credit limit.

   If the client's workload is high, consider upgrading to a higher plan or adjusting the workflow to be more efficient.

8. 08

   Document the agent configuration and deliver a handoff guide to the client.

   Include the prompt used, output format, schedule, and how to pause or modify the agent.