Customer Sentiment Analysis AI

What is Customer Sentiment Analysis AI?

Customer Sentiment Analysis AI is an all-in-one Apify Actor that scrapes customer reviews from Facebook, Google Maps, Booking.com, and Amazon in parallel, then runs AI-powered sentiment analysis on every review using your own Claude or OpenAI API key. In a single click, you get an aggregate sentiment report with scores, trends, themes, and competitor comparisons - plus individually scored reviews ready for dashboards, spreadsheets, or BI tools.

No juggling multiple scrapers and separate analysis pipelines. This Actor handles the entire workflow: collect, analyze, and summarize customer feedback from four major platforms at once.

Features - what can Customer Sentiment Analysis AI do?

• Multi-platform review scraping - Collects reviews from Facebook pages, Google Maps places, Booking.com hotels, and Amazon products in a single run
• AI sentiment analysis - Scores every review (1-10) and labels it positive, neutral, or negative using Claude or GPT models
• Aggregate sentiment report - Executive summary, overall sentiment score, top positive/negative themes, monthly trends, and actionable recommendations
• Competitor sentiment comparison - Supply multiple entities and get a side-by-side breakdown of how each brand, hotel, or product is perceived
• Bring Your Own Key (BYOK) - Uses your own Anthropic or OpenAI API key, so there are no hidden AI costs baked into the Actor price
• Trend analysis - Monthly sentiment trends reveal whether customer perception is improving or declining over time
• Parallel scraping - All four platforms are scraped simultaneously, dramatically reducing total run time
• Flexible time filters - Analyze reviews from the last 7 days up to all time
• Export anywhere - Results integrate with Apify's ecosystem: export to Google Sheets, Slack, Zapier, webhooks, or download as JSON/CSV/Excel

What data can you extract?

Aggregate report fields

Field	Description
overallSentimentScore	Combined sentiment score across all sources (1-10)
overallSentimentLabel	Human-readable label: positive, neutral, or negative
executiveSummary	AI-generated summary of customer sentiment across all reviews
totalReviewsAnalyzed	Total number of reviews processed
topPositiveThemes	Most frequently mentioned positive themes (e.g., "friendly staff", "fast delivery")
topNegativeThemes	Most frequently mentioned negative themes (e.g., "slow response", "quality issues")
recommendations	AI-generated actionable recommendations based on review patterns
sentimentTrend	Monthly sentiment score breakdown showing trajectory over time
competitorComparison	Side-by-side sentiment comparison when multiple entities are provided
sourceBreakdown	Per-platform metrics: review count, average sentiment, and top themes for each source

Individual review fields

Field	Description
source	Platform the review came from (Facebook, Google Maps, Booking.com, or Amazon)
entity	The specific business, place, hotel, or product being reviewed
text	Full review text
date	Date the review was posted
rating	Original star/number rating from the platform
sentimentScore	AI-assigned sentiment score (1-10)
sentimentLabel	Sentiment classification: positive, neutral, or negative
themes	Array of themes detected in the review (e.g., ["customer service", "pricing"])

How to run customer sentiment analysis

1. Go to the Actor page on Apify Store and click Start.
2. Add your review URLs - Paste Facebook page URLs, Google Maps place URLs, Booking.com hotel URLs, and/or Amazon product URLs into the corresponding input fields.
3. Choose a time period - Select how far back to analyze: last 7 days, 30 days, 90 days, 6 months, 1 year, or all time.
4. Set the review limit - Choose the maximum number of reviews to collect per source (default is 100).
5. Select your LLM provider - Pick either Anthropic Claude or OpenAI GPT, then paste your API key.
6. Click Run - The Actor scrapes all platforms in parallel, analyzes every review with AI, and delivers both the aggregate report and individually scored reviews.
7. Export your results - Download as JSON, CSV, or Excel; push to Google Sheets; or connect via Apify API and webhooks.

Using the Apify API

You can also run customer sentiment analysis programmatically via the Apify API. Call the Actor endpoint with your input JSON, then poll or use a webhook to retrieve results. This is ideal for automated reputation monitoring or integrating sentiment data into your own dashboards.

How much does it cost?

The Actor's cost has two components:

• Apify platform compute - You pay for the compute resources used during scraping and orchestration. A typical run analyzing 100 reviews from two platforms costs approximately $0.10-$0.50 in Apify credits depending on the number of sources and reviews. Check the Apify pricing page for current rates.
• LLM API costs (your own key) - Sentiment analysis runs on your own Claude or OpenAI API key. Costs depend on the model and review volume. Analyzing 100 reviews with Claude claude-sonnet-4-20250514 costs roughly $0.05-$0.15; GPT-4o is comparable. Because you bring your own key, there are no hidden AI markups in the Actor's pricing.

Tip: Start with a small maxReviewsPerSource value (e.g., 20) to estimate costs before scaling up.

Input fields

Field	Type	Description
facebookUrls	String list	Facebook page URLs to scrape reviews from
googleMapsUrls	String list	Google Maps place URLs to scrape reviews from
bookingUrls	String list	Booking.com hotel/property URLs to scrape reviews from
amazonUrls	String list	Amazon product URLs to scrape reviews from
timePeriod	Select	Time range filter: last 7 days, 30 days, 90 days, 6 months, 1 year, or all time
maxReviewsPerSource	Integer	Maximum number of reviews to collect per source (default: 100)
llmProvider	Select	AI provider: Anthropic Claude or OpenAI GPT
llmApiKey	Secret	Your own API key for the selected LLM provider
llmModel	String (optional)	Override the default model. Defaults: claude-sonnet-4-20250514 for Anthropic, gpt-4o for OpenAI

Output examples

Aggregate report

{
    "overallSentimentScore": 7.4,
    "overallSentimentLabel": "positive",
    "executiveSummary": "Customer sentiment across 312 reviews is predominantly positive. Guests consistently praise the central location and friendly staff, while the most common complaints involve slow Wi-Fi and limited breakfast options. Sentiment has trended upward over the past 3 months following recent renovations.",
    "totalReviewsAnalyzed": 312,
    "topPositiveThemes": ["central location", "friendly staff", "clean rooms", "good value"],
    "topNegativeThemes": ["slow Wi-Fi", "limited breakfast", "street noise", "small bathrooms"],
    "recommendations": [
        "Upgrade Wi-Fi infrastructure - mentioned negatively in 23% of reviews",
        "Expand breakfast menu variety - recurring complaint across Google Maps and Booking.com",
        "Add soundproofing to street-facing rooms to address noise complaints"
    ],
    "sentimentTrend": [
        { "month": "2026-01", "averageScore": 6.8 },
        { "month": "2026-02", "averageScore": 7.1 },
        { "month": "2026-03", "averageScore": 7.4 },
        { "month": "2026-04", "averageScore": 7.6 }
    ],
    "competitorComparison": [
        { "entity": "Hotel Aurora", "sentimentScore": 7.4, "reviewCount": 156 },
        { "entity": "Hotel Belvedere", "sentimentScore": 6.9, "reviewCount": 156 }
    ],
    "sourceBreakdown": [
        { "source": "Google Maps", "reviewCount": 120, "averageScore": 7.6 },
        { "source": "Booking.com", "reviewCount": 98, "averageScore": 7.2 },
        { "source": "Facebook", "reviewCount": 54, "averageScore": 7.1 },
        { "source": "Amazon", "reviewCount": 40, "averageScore": 7.5 }
    ]
}

Individual review

{
    "source": "Google Maps",
    "entity": "Hotel Aurora",
    "text": "Fantastic location right in the city center. Staff were incredibly helpful and recommended great restaurants. Only downside was the Wi-Fi kept dropping during video calls.",
    "date": "2026-03-15",
    "rating": 4,
    "sentimentScore": 7,
    "sentimentLabel": "positive",
    "themes": ["central location", "friendly staff", "restaurant recommendations", "slow Wi-Fi"]
}

Use cases for consumer sentiment analysis

• Hotel and restaurant review analysis - Aggregate reviews from Google Maps and Booking.com to understand guest satisfaction, spot recurring complaints, and benchmark against nearby competitors
• Product feedback analysis - Collect Amazon reviews for your products (or competitors') to identify feature requests, quality issues, and what customers love most
• Brand reputation monitoring - Track sentiment across Facebook and Google Maps over time to detect PR crises early and measure the impact of marketing campaigns
• Competitor sentiment comparison - Input URLs for multiple competing businesses or products and get a side-by-side sentiment breakdown with scores, themes, and trends
• Customer experience improvement - Use AI-generated recommendations and theme analysis to prioritize operational changes that will have the biggest impact on customer satisfaction
• Marketing strategy from sentiment trends - Identify what customers praise most and use those themes in ad copy, landing pages, and social proof

Frequently asked questions

Which platforms does Customer Sentiment Analysis AI support?

The Actor currently supports four platforms: Facebook pages, Google Maps places, Booking.com hotels, and Amazon products. You can scrape any combination of these in a single run - for example, just Google Maps and Booking.com for a hotel, or just Amazon for a product.

Do I need my own API key for the AI analysis?

Yes. Customer Sentiment Analysis AI uses a Bring Your Own Key (BYOK) model. You provide your own Anthropic Claude or OpenAI GPT API key. This means you pay the LLM provider directly at their standard rates with no markup. The Actor itself only charges for Apify compute used during scraping.

How accurate is the AI sentiment analysis?

Accuracy depends on the LLM model you choose. Both Claude claude-sonnet-4-20250514 and GPT-4o deliver high-quality sentiment scoring on customer reviews, with strong theme detection and nuanced understanding of context, sarcasm, and mixed opinions. The 1-10 scoring scale provides more granularity than simple positive/negative classifiers. For best results, use the default model recommendations.

How does this compare to dedicated sentiment analysis tools?

Most review sentiment analysis tools require you to scrape reviews separately and then upload them. This Actor combines scraping and analysis in one step - no data pipelines, no CSV uploads, no separate tools. It also supports four platforms simultaneously with parallel scraping and includes competitor comparison and trend analysis out of the box. The BYOK model means you use state-of-the-art LLMs (Claude or GPT) instead of simpler NLP classifiers.

Can I use this Actor via the Apify API for automated monitoring?

Absolutely. You can schedule the Actor to run daily, weekly, or monthly via the Apify platform, and use webhooks or the Apify API to push results to your dashboards, Slack channels, Google Sheets, or data warehouses. This makes it ideal for ongoing brand sentiment monitoring and customer feedback analysis at scale.