Agent Builders Evaluation Gate (Onboarding)

Agent Builders Evaluation Gate (Onboarding)

1. 01

   Define the named client workflow the agent will serve

   Pick one concrete process, such as lead qualification or report compilation, and document its current inputs, outputs, and decision points before evaluating any platform.

2. 02

   Map required controls to the workflow

   List the governance, testing, and audit requirements, including how the agent's actions will be logged and approved, since the defensible value lies in workflow ownership around the shell.

3. 03

   Assess white-label and resale capabilities

   Check if the platform supports custom domains, branding, and client-facing deployment, as agencies need to productize the agent under their own brand to build retainer value.

4. 04

   Estimate implementation effort in hours

   Calculate the time to configure flows, memory, tool calls, and channel routing, and compare it against the agency's available delivery capacity.

5. 05

   Review the commercial model and pricing

   Determine whether the platform charges per seat, per agent, or per usage, and model the margin the agency can achieve when reselling the agent to clients.

6. 06

   Run a pilot on a low-risk task

   Test the agent on a repeatable internal task, such as competitor monitoring, and document time saved and output quality before committing to a client deployment.

7. 07

   Document the evaluation against the named workflow

   Record which platform met the controls, effort, and commercial criteria, and keep this as a reference for future client proposals.