The AI Review Insights Advanced Dashboard lets you transform customer reviews into actionable insights for your business.
You receive a large volume of reviews every day. Behind this volume lies a wealth of information that is often underutilized. Understanding what consumers think is crucial for improving the customer experience, prioritizing product development, or adjusting your marketing strategy.
AI Review Insights automates large-scale analysis of all your consumer reviews by identifying the key themes discussed (reception, delivery, quality, price, etc.) and detecting associated sentiments with a high level of accuracy.
This gives you an immediate understanding of expectations, pain points, and weak signals, so you can take action more quickly!
To access this dashboard, you must be on our Advanced Pricing Plan. We invite you to contact us to learn more.
Access the Dashboard
To
access this dashboard, go to the left sidebar menu
Statistics,
then
AI Review Insights
From this dashboard you can access:
- Automatic detection of themes by AI: the algorithm scans all reviews and detects keywords related to the online/in-store purchase experience (reception, friendliness, etc.) and product attributes (quality, price, etc.).
- Advanced sentiment analysis: each theme is enriched with a sentiment score (positive, negative, neutral) with over 95% accuracy.
- Granularity and depth of analysis: analyze your insights by region, product, store, or timeframe to refine your decisions locally or globally.
How does it work?
- Identification and structuring of themes: the system identifies the most frequently mentioned topics in each sector, then groups and normalizes them to create a coherent and cross-functional classification plan.
- Model training: a supervised NLP model is trained to automatically assign themes to review texts, and another to analyze sentiment, using cross-validation.
- Advanced quality control: models are calibrated and evaluated, and a double-check mechanism via LLM ensures result reliability.
You may notice some data differences between this dashboard and the Review Publication dashboard for Product Reviews. Here’s why:
- Updates: AI Review Insights is refreshed once a week, on Mondays. To compare data, we recommend using the most recent full update date.
- Review status: AI Review Insights contains all collected reviews, including rejected reviews. To compare with published reviews, filter the dashboard by review status (published and published after moderation).
-
Product category: Your product category information is provided at the time of the consumer’s order/purchase experience. If some orders do not include a product category, AI Review Insights will automatically assign it based on other orders containing the same product.
Unlike the Review Publication dashboard, which shows raw data, AI Review Insights intelligently reconstructs product categories, which can sometimes create slight differences.
Shopping Experience
How to analyze it?
This section is based on the analysis of brand reviews left by your consumers. It allows you to explore feedback related to the purchase experience across multiple dimensions.
Refine your analysis using the following filters: timeframe (dates), purchase channel, website, country, and locations (if these are configured on your account).
You can also compare results with a previous period.
Displayed data
- Number of brand reviews published during the selected period, with a trend (up or down compared to the previous period).
- Sentiment analysis by detected themes. For example: overall satisfaction, digital experience, responsiveness, team expertise, reception and friendliness, choice, etc.
For each detected theme, the following will be available:
- The number of reviews concerned
- The percentage of mentions (e.g., friendliness mentioned in 20% of reviews)
- The associated sentiment (positive, neutral, or negative) and its trend (improving or declining)
Example: An online detergent brand finds that its customers frequently mention the ease of ordering and subscription management, but with rather negative feedback: the checkout process is deemed too complex, unclear, or inflexible in managing subscriptions.
In response, the brand simplifies its purchase funnel and implements an automated renewal subscription that is more flexible and convenient, preventing customers from running out of detergent. Customer feedback on these topics then becomes more positive and consistent.
Products
This section is based on analysis of product reviews left by your consumers. It allows you to explore feedback related to product satisfaction across multiple dimensions.
Refine your analysis using the following filters: timeframe (dates), purchase channel, website, country, and locations (if these are configured on your account). You can also compare results with a previous period.
Displayed data
- Number of product reviews published during the selected period, with a trend (up or down compared to the previous period).
- Sentiment analysis by detected themes. For example: quality, usability, price, aesthetics, color, etc.
For each detected theme, the following will be available:
- The number of reviews concerned
- The percentage of mentions (e.g., quality mentioned in 20% of reviews)
- The associated sentiment (positive, neutral, or negative) and its trend (improving or declining)
Example: A sofa brand notices that fabric quality is frequently mentioned in reviews, with a rising negative sentiment, mainly due to the fabric pilling quickly. They then decide to review material choices by selecting more wear-resistant fabric and implement more rigorous quality tests to ensure upholstery durability before production.
Thematics
This section allows you to select a specific theme to deepen the analysis of consumer feedback.
Example theme: Choice (Purchase Experience category)
This theme groups reviews related to offer diversity, relevance, and availability.
Displayed data
- Positive sentiment rate (%), compared to the previous period, with trend indication (up or down).
- Evolution of positive sentiment over the last 12 months, on a weekly or monthly view, compared to the previous year.
- Number of reviews mentioning the theme.
- Or proportion of reviews (e.g., 20% of reviews mention the theme):
Sentiment by website/country
- Number of reviews by site
- Percentage of mentions
- Associated sentiment (positive, neutral, or negative) and its trend
Detailed list of associated reviews
For each review mentioning the selected theme, the following is displayed:
- The rating
- The publication date
- The review comment
- The country
- The website
- The associated sentiment
Example: A cosmetics brand finds that the range of shades is frequently mentioned in reviews, with a particularly negative sentiment notably in Spain. They decide to expand their range to better cover different skin types and adapt their offering by country.
They also launch a campaign showcasing the diversity of their products to meet the expressed expectations.