Artificial Intelligence Policy

1. Overview

MuniGrowth uses large language models (accessed via OpenRouter and related providers) to assist municipal teams with LED planning, grant discovery, document drafting, and coaching workflows. This policy explains capabilities, limits, and your obligations under evolving regulations including the EU AI Act framework for transparency and human oversight.

2. How AI Is Used

• LED Coach — guided questions for needs assessment, SWOT, goals, and projects;
• Grant matching — scoring and narrative suggestions based on project and donor data;
• Document generation — drafts for plans, dossiers, and application sections;
• AI Strategy Engine — conversational analysis using your tenant context snapshot.

3. What AI Cannot Do

• Guarantee grant approval or funding outcomes;
• Replace legal, financial, or engineering sign-off;
• Access data from other municipalities;
• Provide real-time official statistics without verification;
• Operate without human review on consequential submissions.

4. User Responsibilities

• Review all AI outputs before council, donor, or public submission;
• Correct factual errors and cite official sources where required;
• Do not enter special-category personal data unless legally justified and approved by your DPO;
• Assign accountable officials for AI-assisted deliverables.

5. Reporting AI Errors

Report harmful, biased, or materially incorrect outputs to contact@munigrowth.com with tenant name, feature (Coach, Strategy, etc.), date, and a description. We use reports to tune prompts and monitoring.

6. Data Used for AI

Prompts may include your municipality context (plans, projects, grants) to personalize responses. We do not use your identifiable content to train public foundation models. Aggregated, de-identified patterns may inform product quality improvements as described in our Privacy Policy.