Automating Google review evaluation for Libeřské lahůdky saves dozens of hours per month

Libeřské lahůdky offers its delicacies in 21 shops not only in Prague, but also in České Budějovice, for example, and new branches are regularly being added. With the growth in the number of branches comes an enormous increase in customer reviews on Google. Every month, dozens to hundreds of reviews were received, which were not clearly processed anywhere. Processing each of these reviews, i.e. studying them, writing responses and, if necessary, forwarding them to the appropriate colleagues for resolution, took employees up to several tens of minutes, amounting to several tens of hours of work per month.

For Libeřské lahůdky, it was crucial to find a solution that would prevent employees from wasting time manually monitoring reviews and instead allow them to focus only on those that truly required action, such as negative feedback or customer questions. The usual approach of checking the Google account and replying manually was time-consuming and, in the context of the company’s expansion, unsustainable in the long term.

Our solution: automation and intelligent classification

To build a functional system capable of retrieving reviews from all branches across the Czech Republic in real time, it was necessary to gain access to the Google Reviews API. Libeřské lahůdky could not obtain this access directly, as Google grants it only to large multinational companies or certified agencies. At TRITON IT, we hold this certification, which allowed us to retrieve all reviews from every branch and begin processing and classifying them.

Review classification was a key component of the solution. Using n8n, we designed a workflow that retrieves reviews on a daily basis and analyzes them according to the following criteria:

1. Sentiment – positive or negative
2. Specificity – whether the review describes a concrete experience or is generic
3. Personal vs. generic – whether the review contains a name or personal experience, or is neutral
4. Contains a question – whether the customer asked a question that requires a response

Based on this classification, we defined processing scenarios:

• Positive, generic reviews – the system automatically generates a thank-you message using generative AI
• Reviews containing questions – the system forwards the review to the appropriate employee, who can reply manually or approve an automatically generated personalized response
• Ratings below 5 stars* – the system forwards the review to responsible staff so the case can be handled individually

Implementation in practice

In practice, the system works as follows:

• Daily review retrieval – a bot downloads new reviews every day and classifies them according to the defined criteria
• Forwarding reviews to employees – if action is required, employees receive a link to the specific review with the option to respond
• Automated responses – for routine reviews, the system generates a personalized response, for example: “Thank you, Marek, for your review!”

General reviews that do not require human intervention and can be handled with automated responses account for up to 75% of all reviews. Thanks to this, Libeřské lahůdky can focus only on feedback that truly requires attention, while customer communication remains fast and consistent.

Benefits and impact

By implementing the automated system, Libeřské lahůdky achieved several key benefits:

• Savings of dozens of hours per month
• Faster responses to customer questions and negative reviews
• Better visibility into customer feedback
• Consistent communication tone and quality

In addition, we helped Libeřské lahůdky define internal processes and trained employees to use the system effectively, without losing control over customer communication. We also create clear dashboards from incoming reviews, providing ongoing insight into customer satisfaction at individual branches.

The Libeřské lahůdky case demonstrates how a combination of data access, automation, and intelligent classification can streamline daily operations and significantly improve the customer experience. It is not just about technology, but also about well-designed workflows and clearly defined processes.