Mistral Prompt for Analyzing Customer Reviews
Mistral, the AI model developed by Mistral AI, excels at processing and analyzing French-language text data. Customer review analysis is a strategic lever for any company looking to understand the perception of its products or services. Thanks to its nuanced understanding of natural language, Mistral can identify dominant sentiments, extract recurring themes, and detect weak signals within large volumes of customer feedback. Whether you manage an e-commerce site, a SaaS platform, or a B2B service, automating your review analysis saves hours of manual work while providing more precise, structured insights. This prompt is designed to transform a raw corpus of reviews into an actionable analysis report, with concrete recommendations prioritized. It works equally well with Google, Trustpilot, Amazon reviews, or any other source of customer feedback.
Paste in your AI
Paste this prompt in ChatGPT, Claude or Gemini and customize the variables in brackets.
You are an analyst specializing in customer experience and sentiment analysis. I will provide you with a set of customer reviews. Analyze them following this structured methodology:
-
Overall Sentiment Analysis: Classify each review as positive, neutral, or negative. Provide the percentage breakdown.
-
Recurring Themes: Identify the 5 to 8 main themes mentioned (e.g., product quality, customer service, delivery, value for money, ease of use). For each theme, indicate the dominant sentiment and the number of mentions.
-
Strengths: List the 3 elements most appreciated by customers, with representative exact quotes.
-
Friction Points: List the 3 main sources of dissatisfaction, with exact quotes and a severity rating (low / medium / critical).
-
Weak Signals: Identify 2 to 3 emerging trends or isolated mentions that deserve attention.
-
Estimated NPS Score: Based on the tone and content of the reviews, estimate an indicative Net Promoter Score.
-
Actionable Recommendations: Propose 5 concrete actions ranked by impact (high/medium) and effort (low/medium/high), in a matrix format.
Output format: structured report with headings, bullet points, and Markdown tables.
Here are the reviews to analyze:
[PASTE_YOUR_REVIEWS_HERE]
Personalize this prompt with Léa
Answer 3 questions and Léa tailors the prompt to your situation.
Why this prompt works
This prompt uses a methodical 7-step breakdown that forces the model to produce an exhaustive analysis rather than a superficial summary. The requirement for exact quotes anchors the analysis in real data and avoids hallucinations. The concluding impact/effort matrix transforms the analysis into an action plan directly usable by product and marketing teams.
Use Cases
Variants
Expected Output
You get a structured Markdown report including the percentage breakdown of sentiments, a table of recurring themes with their frequency and tone, key positive and negative verbatims, and a matrix of 5 prioritized recommendations. The report is directly shareable with your teams and usable to guide your product, customer service, or communication decisions.
Frequently Asked Questions
How many reviews can I analyze at once with Mistral?
The limit depends on the context window of the Mistral model you're using. Mistral Large supports up to 128K tokens, which is roughly 300 to 500 average-sized reviews in a single prompt. For larger volumes, split your reviews into batches of 200–300 and request a consolidated summary at the end. Tip: number your reviews to make referencing quotes in the analysis easier.
Can Mistral analyze reviews in multiple languages at once?
Yes, Mistral handles multilingual content very well, including French, English, Spanish, German, and Italian. You can submit a mixed-language corpus and explicitly ask in your prompt for the analysis to be delivered in English. Simply specify: 'The reviews are in multiple languages, provide the entire analysis in English.' The model will automatically detect each review's language.
How can I get more reliable sentiment analysis results?
To improve accuracy, add context to your prompt: specify your industry, the type of product or service involved, and your rating scale if applicable. For example, a 3-star out of 5 review in hospitality doesn't carry the same meaning as in e-commerce. You can also ask Mistral to assign a confidence score (low/medium/high) to each sentiment classification to flag ambiguous reviews that need human review.
Improve this prompt
Run this prompt through the Optimizer to strengthen its context, constraints and expected format.
Improve this prompt with the OptimizerComments
Be the first to comment on this prompt.
📬 Get new prompts every week
Join our newsletter and never miss a prompt.