Explain the AI safety-check workflow

Co-Authored-By: codex <codex@openai.com>

static/index.html

@@ -30,7 +30,7 @@
         <p class="hero-text">Check suspicious bills, bank alerts, FBR-style messages, challans, courier notices, and SMS screenshots for common scam signals.</p>
         <div class="trust-note">
           <span class="shield" aria-hidden="true">✓</span>
-          <p><strong>Model-powered analysis.</strong> Your input is sent to the configured Modal endpoint and is not saved by this app.</p>
+          <p><strong>AI-assisted safety check.</strong> The AI reads the notice, identifies scam signals, and returns a structured risk assessment with safer next steps.</p>
         </div>
       </div>
       <div class="hero-image">