Monitor new questionnaire responses in Google Sheets, evaluate them with Azure GPT-4o-mini, combine with existing data, and update the candidate dataset in near real-time, delivering structured JSON takeaways.
The AI agent automatically scores candidate questionnaire responses using Azure OpenAI GPT-4o-mini and returns a structured score plus takeaways. It looks up existing candidate data from a central Google Sheet to fuse with the questionnaire results and compute a final score. It parses and normalizes outputs, supports no-code customization of questions and weights, and writes updates back to the candidate store in near real-time.
End-to-end scoring and data fusion to support hiring decisions.
Monitor new questionnaire submissions in Google Sheets every minute.
Evaluate responses with Azure OpenAI GPT-4o-mini to produce a score (0–30) and takeaways.
Parse and normalize JSON output safely, handling code fences and errors.
Lookup existing candidate data in the Resume store to fuse with questionnaire results.
Compute Final Score = Existing Score + Questionnaire Score.
Update or append records by candidate name while preserving existing data.
Before: manual scoring across questionnaires with inconsistent criteria and no audit trail. After: automated, structured scoring with real-time updates, centralized data, and auditable takeaways.
A simple 3-step system you can use without technical expertise.
Poll Google Sheets every minute for new questionnaire submissions and fetch candidate identifiers.
Send responses to Azure OpenAI GPT-4o-mini to generate a structured JSON with a score (0–30) and takeaways, then parse and normalize the results.
Lookup existing candidate data in the Resume store, compute Final Score, and update or append the candidate's row, preserving previous data.
A realistic scenario showing time-to-score and the resulting update.
9:12 AM: A new questionnaire response arrives in BD Questionarie Form Responses 1. Within ~60 seconds, the AI agent evaluates the responses with GPT-4o-mini, returning a score of 23 and key takeaways. The agent then looks up the candidate in the Resume store, which shows an existing score of 15. The Final Score becomes 38, and the agent updates the candidate's row in Resume store with the new score and notes. The hiring team can access the updated profile in real time for faster decision-making.
Roles involved in evaluation and decision-making can leverage this AI agent.
Receives immediately actionable scores and takeaways for faster decision-making.
Gets auditable scoring data and consolidated candidate profiles for reports.
Automates data fusion and updates to the central Google Sheet.
Tracks candidate progress with consistent scoring across teams.
Can customize questions and weights without writing code.
Ensures data alignment across systems and stakeholders.
The AI agent plugs into your data and workflow tools.
Reads new responses from BD Questionarie / Form Responses 1 and matches candidates to existing profiles.
Fetches existing candidate data to fuse with questionnaire results and update scores by name.
Evaluates responses and returns a structured JSON with score and takeaways.
Orchestrates polling, credentials, and data routing across Sheets and OpenAI calls.
Common scenarios where this AI agent adds value.
Practical answers to common questions.
It reads from the BD Questionarie Google Sheet (Form Responses 1) for new submissions and from the Resume store Google Sheet (Sheet2) for existing candidate data. It uses n8n to securely manage credentials and data routing between these sources and Azure OpenAI. The agent never exposes raw candidate responses beyond the structured JSON it returns, and updates are written back to the central Resume store. All data handling adheres to the configured access controls in your n8n setup.
Yes. The workflow includes configurable evaluation questions and adjustable weights. You can modify question content, scoring ranges, and how different responses contribute to the final score. Changes apply to new submissions and retroactive scoring when re-run, depending on your workflow configuration. This keeps the scoring aligned with your current hiring criteria.
The system includes resilient JSON parsing that handles code fences and common formatting errors. If parsing fails, the agent logs the error, returns a fallback score (e.g., 0) and the takeaways field, and alerts you for manual review. It then stores the parsed output in a safe, auditable location for debugging. You can adjust parsing rules to reduce failure rates.
The integration relies on secure credentials managed by n8n and standard Google Sheets access permissions. Data passed to Azure OpenAI is limited to the structured content required for scoring and takeaways, not raw long-form responses. You can enable additional masking or route sensitive fields through separate, restricted flows. Always align usage with your organization’s data policies and vendor terms.
The polling cadence is near real-time, typically every minute for new questionnaire submissions. The entire cycle from submission to final record update is designed to complete within tens of seconds in typical conditions. Volume, network latency, and API limits can affect latency, but the system is built to handle bursts with graceful backoff. You can adjust the polling interval in your n8n workflow.
Yes. The agent updates candidate records in a way that preserves prior data, allowing you to view historical scores and takeaways. You can implement versioning in the Resume store to track changes over time. If needed, you can revert to a previous score by restoring a prior row state. Audit trails are maintained via logs and JSON outputs.
If no matching profile exists, the agent creates a new record in the Resume store with the questionnaire score and takeaways. It then initializes a base score and links to the candidate’s identifier. This ensures new applicants are included in the centralized dataset from their first submission. Subsequent updates will merge with this new profile as more data becomes available.
Monitor new questionnaire responses in Google Sheets, evaluate them with Azure GPT-4o-mini, combine with existing data, and update the candidate dataset in near real-time, delivering structured JSON takeaways.
