AI Automation
AI Brand Visibility Monitor
Task decomposition
- User submits a form with product category, target brand, number of brands to recommend, and number of test runs.
- Workflow generates one identical test prompt per run.
- Each prompt runs through a loop and is sent to a recommendation agent, which returns a structured list of brands.
- Each response is parsed to check whether the target brand appears and at what position.
- Once all runs are processed, results are aggregated into visibility and top-position rates, plus the top competing brands.
- A second agent drafts a full content-update report from the aggregated results.
- Report is converted to HTML and returned as a completion screen.
Objective of the workflow
- Measures how consistently and favorably a brand gets recommended by AI answer engines, the emerging equivalent of search ranking.
- Repeats the same prompt many times and scores the pattern instead of relying on a single spot-check.
Architecture
