Project Background

The client, a sustainability consultancy operating in the UK and France, had a manual and highly inefficient process for completing sustainability questionnaires, involving over 130 questions and referencing extensive documentation (300+ pages). The primary objectives were to automate the questionnaire process, accelerate completion time, and ensure consistent, accurate scoring.

Challenges

The client faced the challenge of manually completing extensive sustainability questionnaires, which was extremely time-consuming, often taking up to 3 days per questionnaire. The complexity and length of scoring rubrics also led to frequent human errors, affecting reliability and consistency.

Constraints included:

• The urgent requirement to speed up questionnaire processing to meet strict client deadlines.
• Mandatory integration with existing Microsoft Word and Excel workflows.
• The necessity of significantly reducing costs per questionnaire through effective automation.

Solution

To address the client's needs, we developed an AI-powered "Questionnaire Agent" using the Agency Swarm framework. This agent automated document parsing, rubric-based scoring, and rapid Excel report generation. The solution featured a user-friendly Gradio interface for seamless interaction.

Key Features and Functionality

• Automated parsing and analysis of extensive sustainability documentation.
• Accurate rubric-based scoring to ensure consistency.
• Rapid generation of detailed Excel reports.
• Integration with Google Drive for secure storage and ease of file management.

Results and Impact

The automation drastically reduced questionnaire processing time from days to approximately 20 minutes per submission. Operational costs per questionnaire dropped significantly, averaging $1.50–$2.50. Accuracy and consistency were substantially improved, positively impacting client satisfaction.

Client feedback emphasized the transformational impact: "This solution has revolutionized our workflow, significantly improving our efficiency and consistency."

Metrics demonstrated significant improvements, including a substantial reduction in processing time, cost savings per questionnaire, and enhanced accuracy.

Lessons Learned

• Automation of complex documentation processes significantly enhanced efficiency and accuracy.
• Structured rubric application minimized human error.
• Continuous updates and attention to scalability were crucial for long-term success.
• Future projects could benefit from expanding the solution to other document types and increasing user interaction capabilities.

Conclusion

This AI-powered solution drastically improved the client’s questionnaire processing, significantly reducing manual workload, costs, and errors. Future plans include extending the solution to other compliance-related documents and enhancing the scalability and user interaction capabilities.