Benjamin Nwandu.Let’s talk
All projects

Control & reliability

Planned

Human-in-the-Loop Automation

Let AI prepare the action. Keep consequential decisions with a human.

01 / THE PROBLEM

Why this matters.

Automation can create avoidable risk when proposed refunds, account changes, or sensitive updates execute without appropriate oversight.

02 / THE DIRECTION

The proposed solution.

Evaluate action risk, execute only permitted low-risk work, and route higher-risk proposals through an explicit approval or rejection checkpoint with an audit trail.

03 / ARCHITECTURE

How the pieces connect.

Conceptual flow — proposed architecture

  1. 01Business trigger
  2. 02Proposed action
  3. 03Risk evaluation
  4. 04Human approval
  5. 05Execution
  6. 06Audit trail

04 / ENGINEERING DECISIONS

Designed with intent.

  • Bind approval to the specific proposed action.
  • Make rejection a first-class outcome.
  • Keep risk rules and permissions deterministic.

05 / EVALUATION APPROACH

What needs to hold true.

  • Verify rejected actions never execute.
  • Test missing or expired approval.
  • Check that changes to an action require renewed approval.

Technology direction.

n8n · plannedClaude · plannedWebhooks · plannedBusiness APIs · planned

Let’s build something useful

Good conversations.
Better systems.

Have a conversational AI challenge, an automation idea, or a role in mind?