diff --git a/src/components/AIQueryInterface.astro b/src/components/AIQueryInterface.astro
index 01a934b..5e3f1b2 100644
--- a/src/components/AIQueryInterface.astro
+++ b/src/components/AIQueryInterface.astro
@@ -979,7 +979,7 @@ class AIQueryInterface {
             </svg>
           </div>
 
-          <div style="text-align: right; margin: 0.5rem 0 1rem 0;">
+          <div style="margin: 0.5rem 0 1rem 0;">
             <button id="download-results-btn" class="btn btn-secondary">
               <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" style="margin-right: 0.5rem;">
                 <path d="M21 15v4a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2v-4"/>
@@ -1010,19 +1010,27 @@ class AIQueryInterface {
               </div>
               
               <div class="insights-section">
-                ${stats.qualityMetrics.aiTransparency >= 80 ? `
-                  <div class="insights-header success">✅ Hohe Transparenz</div>
+                ${stats.keyInsights && stats.keyInsights.length > 0 ? `
+                  <div class="insights-header success">✅ Erkenntnisse</div>
+                  <ul class="insights-list">
+                    ${stats.keyInsights.map(insight => `<li>${insight}</li>`).join('')}
+                  </ul>
+                ` : ''}
+                
+                ${stats.potentialIssues && stats.potentialIssues.length > 0 ? `
+                  <div class="insights-header warning">⚠️ Hinweise</div>
+                  <ul class="insights-list">
+                    ${stats.potentialIssues.map(issue => `<li>${issue}</li>`).join('')}
+                  </ul>
+                ` : ''}
+                
+                ${(!stats.keyInsights || stats.keyInsights.length === 0) && (!stats.potentialIssues || stats.potentialIssues.length === 0) ? `
+                  <div class="insights-header">📊 Analyse-Details</div>
                   <ul class="insights-list">
                     <li>${stats.qualityMetrics.aiTransparency}% der Entscheidungen mit nachvollziehbarer Begründung</li>
-                    <li>Durchschnittliche Verarbeitungszeit: ${formatDuration(stats.qualityMetrics.avgProcessingTime)}</li>
+                    <li>Analysequalität: ${this.getQualityDisplayText(stats.analysisQuality)}</li>
                   </ul>
-                ` : `
-                  <div class="insights-header warning">⚠️ Transparenz-Hinweise</div>
-                  <ul class="insights-list">
-                    <li>Nur ${stats.qualityMetrics.aiTransparency}% der Entscheidungen vollständig dokumentiert</li>
-                    <li>Einige KI-Entscheidungen ohne detaillierte Begründung</li>
-                  </ul>
-                `}
+                ` : ''}
               </div>
             </div>
 
@@ -1068,6 +1076,17 @@ class AIQueryInterface {
     }
   }
 
+  getQualityDisplayText(analysisQuality) {
+    const qualityMap = {
+      'excellent': 'Ausgezeichnet',
+      'good': 'Gut', 
+      'fair': 'Zufriedenstellend',
+      'poor': 'Verbesserungswürdig',
+      'unknown': 'Unbekannt'
+    };
+    return qualityMap[analysisQuality] || 'Unbekannt';
+  }
+
   setupUnifiedUploadSystem() {
     const previousUploadInput = document.getElementById('upload-previous-analysis');
     if (previousUploadInput) {
@@ -1113,13 +1132,20 @@ class AIQueryInterface {
         totalTime: 0,
         avgConfidence: 0,
         stepCount: 0,
+        highConfidenceSteps: 0,
+        lowConfidenceSteps: 0,
+        phaseBreakdown: {},
         aiDecisionCount: 0,
         embeddingsUsageCount: 0,
         toolSelectionCount: 0,
         qualityMetrics: {
           avgProcessingTime: 0,
+          confidenceDistribution: { high: 0, medium: 0, low: 0 },
           aiTransparency: 0
-        }
+        },
+        analysisQuality: 'unknown',
+        keyInsights: [],
+        potentialIssues: []
       };
     }
 
@@ -1129,25 +1155,115 @@ class AIQueryInterface {
       ? Math.round(validConfidenceEntries.reduce((sum, entry) => sum + entry.confidence, 0) / validConfidenceEntries.length)
       : 0;
     
+    const highConfidenceSteps = auditTrail.filter(entry => (entry.confidence || 0) >= 80).length;
+    const lowConfidenceSteps = auditTrail.filter(entry => (entry.confidence || 0) < 60).length;
+    const mediumConfidenceSteps = auditTrail.length - highConfidenceSteps - lowConfidenceSteps;
+
+    // Enhanced metrics
     const aiDecisionCount = auditTrail.filter(entry => entry.action === 'ai-decision').length;
     const embeddingsUsageCount = auditTrail.filter(entry => entry.metadata?.embeddingsUsed).length;
     const toolSelectionCount = auditTrail.filter(entry => entry.action === 'selection-decision').length;
-    
+
+    // Phase breakdown
+    const phaseBreakdown = {};
+    auditTrail.forEach(entry => {
+      const phase = entry.phase || 'unknown';
+      if (!phaseBreakdown[phase]) {
+        phaseBreakdown[phase] = { count: 0, avgConfidence: 0, totalTime: 0 };
+      }
+      
+      phaseBreakdown[phase].count++;
+      phaseBreakdown[phase].totalTime += entry.processingTimeMs || 0;
+    });
+
+    // Calculate average confidence per phase
+    Object.keys(phaseBreakdown).forEach(phase => {
+      const phaseEntries = auditTrail.filter(entry => entry.phase === phase);
+      const validEntries = phaseEntries.filter(entry => typeof entry.confidence === 'number');
+      
+      if (validEntries.length > 0) {
+        phaseBreakdown[phase].avgConfidence = Math.round(
+          validEntries.reduce((sum, entry) => sum + entry.confidence, 0) / validEntries.length
+        );
+      }
+    });
+
     const avgProcessingTime = auditTrail.length > 0 ? totalTime / auditTrail.length : 0;
     const aiTransparency = auditTrail.length > 0 ? 
       (auditTrail.filter(entry => entry.metadata?.aiPrompt || entry.metadata?.reasoning).length / auditTrail.length) * 100 : 0;
 
+    // RESTORED: Intelligent Analysis Quality Assessment
+    let analysisQuality;
+    if (avgConfidence >= 85 && lowConfidenceSteps === 0) {
+      analysisQuality = 'excellent';
+    } else if (avgConfidence >= 70 && lowConfidenceSteps <= 1) {
+      analysisQuality = 'good';
+    } else if (avgConfidence >= 60 && lowConfidenceSteps <= 3) {
+      analysisQuality = 'fair';
+    } else {
+      analysisQuality = 'poor';
+    }
+
+    // RESTORED: Intelligent Insights Generation
+    const keyInsights = [];
+    const embeddingsUsed = auditTrail.some(e => e.metadata?.embeddingsUsed);
+    if (embeddingsUsed) {
+      keyInsights.push('Semantische Suche wurde erfolgreich eingesetzt');
+    }
+
+    const toolSelectionEntries = auditTrail.filter(e => e.action === 'selection-decision');
+    if (toolSelectionEntries.length > 0) {
+      const avgSelectionConfidence = toolSelectionEntries.reduce((sum, e) => sum + (e.confidence || 0), 0) / toolSelectionEntries.length;
+      if (avgSelectionConfidence >= 80) {
+        keyInsights.push('Hohe Konfidenz bei der Tool-Auswahl');
+      }
+    }
+
+    if (aiTransparency >= 90) {
+      keyInsights.push('Sehr hohe Transparenz der KI-Entscheidungen');
+    }
+
+    if (highConfidenceSteps > auditTrail.length * 0.7) {
+      keyInsights.push('Mehrheit der Analyseschritte mit hoher Sicherheit');
+    }
+
+    // RESTORED: Automatic Issue Detection (excluding processing time warnings)
+    const potentialIssues = [];
+    if (lowConfidenceSteps > 2) {
+      potentialIssues.push(`${lowConfidenceSteps} Analyseschritte mit niedriger Konfidenz`);
+    }
+
+    if (aiTransparency < 50) {
+      potentialIssues.push('Geringe Transparenz der KI-Entscheidungsfindung');
+    }
+
+    const failedAiDecisions = auditTrail.filter(e => e.action === 'ai-decision' && e.confidence < 50).length;
+    if (failedAiDecisions > 0) {
+      potentialIssues.push(`${failedAiDecisions} KI-Entscheidungen mit sehr niedriger Konfidenz`);
+    }
+
     return {
       totalTime,
       avgConfidence,
       stepCount: auditTrail.length,
+      highConfidenceSteps,
+      lowConfidenceSteps,
+      phaseBreakdown,
       aiDecisionCount,
       embeddingsUsageCount,
       toolSelectionCount,
       qualityMetrics: {
         avgProcessingTime,
+        confidenceDistribution: { 
+          high: highConfidenceSteps, 
+          medium: mediumConfidenceSteps, 
+          low: lowConfidenceSteps 
+        },
         aiTransparency: Math.round(aiTransparency)
-      }
+      },
+      analysisQuality,
+      keyInsights,
+      potentialIssues
     };
   }