Google Business Profile Review MCP Server

/** * Sentiment Analysis Prompt * Analyze review sentiment to tailor response tone and approach */ import { logger } from '../../utils/logger.js'; import type { SentimentAnalysisContext } from '../../types/index.js'; export interface SentimentAnalysisPrompt { name: string; description: string; handler: (context: SentimentAnalysisContext) => Promise<string>; } export function createSentimentAnalysisPrompt(): SentimentAnalysisPrompt { return { name: 'sentiment_analysis', description: 'Analyze customer review sentiment and emotions to guide response strategy', handler: async (context: SentimentAnalysisContext): Promise<string> => { try { logger.info('Generating sentiment analysis prompt', { includeEmotions: context.includeEmotions, includeKeywords: context.includeKeywords }); const { reviewText, includeEmotions = true, includeKeywords = true } = context; let prompt = `You are a sentiment analysis expert tasked with analyzing a customer review to understand the emotional context and guide the appropriate response strategy. **REVIEW TO ANALYZE:** "${reviewText}" **YOUR ANALYSIS TASK:** 1. **Overall Sentiment Classification:** - Determine if the sentiment is: POSITIVE, NEGATIVE, NEUTRAL, or MIXED - Provide a confidence score (0-100%) for your classification - If MIXED, explain the positive and negative elements 2. **Sentiment Intensity:** - Rate the intensity on a scale of 1-5: * 1 = Mild/Subtle sentiment * 2 = Moderate sentiment * 3 = Strong sentiment * 4 = Very strong sentiment * 5 = Extreme sentiment (ecstatic or furious) 3. **Primary Concerns/Highlights:** - Identify the main topics or aspects mentioned - Determine which aspects are viewed positively vs negatively - Highlight any specific issues that need addressing 4. **Customer Intent & Expectations:** - What was the customer hoping to achieve with this review? - Are they seeking resolution, recognition, or just sharing experience? - Do they seem open to future engagement with the business?`; if (includeEmotions) { prompt += ` 5. **Emotional Analysis:** - Identify specific emotions present (joy, frustration, disappointment, appreciation, anger, surprise, etc.) - Determine the emotional priority (which emotion is strongest) - Assess the customer's emotional state when writing the review - Note any emotional triggers that should be handled carefully in the response`; } if (includeKeywords) { prompt += ` 6. **Key Phrases & Language Patterns:** - Extract important keywords and phrases used by the customer - Identify language that indicates strong feelings (superlatives, exclamations, etc.) - Note any specific terminology that should be echoed in the response - Highlight any positive language that can be reinforced`; } prompt += ` 7. **Response Strategy Recommendations:** - Suggest the most appropriate response tone (professional, friendly, apologetic, grateful, etc.) - Recommend key points that MUST be addressed in the response - Suggest elements to emphasize or avoid in the reply - Indicate urgency level for response (immediate, prompt, standard) 8. **Risk Assessment:** - Evaluate potential reputation impact (low, medium, high) - Identify any red flags that need special attention - Assess likelihood of escalation if not handled properly - Note if legal or policy concerns are present **OUTPUT FORMAT:** Provide your analysis in a structured format with clear sections for each element above. Be specific and actionable in your recommendations. **ANALYSIS:**`; logger.info('Successfully generated sentiment analysis prompt'); return prompt; } catch (error) { logger.error('Error generating sentiment analysis prompt:', error); throw new Error(`Failed to generate sentiment analysis prompt: ${error instanceof Error ? error.message : 'Unknown error'}`); } } }; }