Best practices applied to your industry Surveys tailored to your needs Demo of our analytics platform A customized quotation
Industry
Channels
BLOG
RateNow, News
RateNow, NewsAI Text Analysis Agent for Customer Comments
Every week, hundreds of comments come in from your surveys. The question is not whether you have enough feedback — it's whether you can read it in time.
Reading and classifying feedback manually does not scale: it consumes hours and lets urgent complaints slip through. RateNow's AI Text Analysis Agent detects the sentiment (positive, negative or neutral) of each comment and classifies it into macro-categories (Emotion, Experience, Staff, Time, Resources) and more specific micro-categories, with an accuracy above 85%. It runs locally, with no data transfer to third parties, and is already active in RateNow accounts.
Diana Bardales, 2026 
7 min 
If you manage satisfaction surveys, you probably recognise the scene: a spreadsheet with thousands of rows, a column of open-ended comments and the feeling that, no matter how much you read, something always slips through. A customer threatening to leave, a patient expressing gratitude for the care received, an employee mentioning a colleague by name. All of it is there, mixed together, and reviewing it manually depends on someone having time that day. The problem is not a lack of feedback — it is the inability to process it at the speed it arrives.
Why reading comments one by one stops working
When survey volume is low, reading comment by comment works reasonably well. The problem arises when the operation grows: multiple touchpoints, different languages, hundreds of responses per day. At that point, manual review has three common consequences:
- Urgent complaints take days to reach the person who should act on them, because they depend on someone detecting them while reading.
- Comments are grouped superficially, without distinguishing whether they refer to staff, waiting times or facilities.
- Recognition for staff (words of thanks, positive interactions) rarely reaches the person mentioned, because no one redirects it.
What the AI Text Analysis Agent does exactly
The AI Text Analysis Agent is the RateNow feature that reads each open-ended comment from a survey and answers two questions: what the person is feeling and exactly what they are talking about. The process follows three steps:
1. It analyses the sentiment
It classifies each comment as positive, negative or neutral, without anyone having to label it manually. A single comment can contain nuances: thanking a staff member and, at the same time, complaining about the waiting time.
2. It classifies into macro-categories and micro-categories
It organises each comment within a general macro-category — Emotion, Experience, Staff, Time or Resources — and a more specific micro-category within it. For example, within Emotion, it distinguishes between gratitude, complaint or opinion.
| Macro-category | What it groups | Example micro-category |
|---|---|---|
| Emotion | The overall tone of the comment | Gratitude, complaint, opinion |
| Experience | Perception of the service, space or facilities | Comfort, cleanliness, accessibility |
| Staff | Interaction with the human team | Professional care, communication |
| Time | Duration of waits or processes | Wait time, service speed |
| Resources | Available material or technical means | Physical space, support technology |
How does it work in practice?
Comment: "Very grateful for the attention from the staff, although I had to wait almost an hour to be seen."
Analysis:
Sentiment = Neutral
Category 1: Emotion | Micro-category 1: Gratitude
Category 2: Time | Micro-category 2: Wait time
Sentiment = Neutral
Category 1: Emotion | Micro-category 1: Gratitude
Category 2: Time | Micro-category 2: Wait time
3. It prioritises and routes
With the sentiment and category already identified, each comment can be directed to whoever needs to see it: a serious complaint to operations, a word of thanks to the person mentioned, a comment about waiting times to the relevant department.
85% accuracy
The AI Text Analysis Agent achieves an accuracy above 85% in comment classification — a level comparable to the judgement of a human analyst — thanks to a model trained specifically to interpret feedback from customers, patients and employees.
Manual reading vs. AI Text Analysis Agent
| Aspect | Manual reading | AI Text Analysis Agent |
|---|---|---|
| Coverage | Depends on the team's available time | 100% of comments, in all languages |
| Speed | Hours or days | Real time |
| Level of detail | General categories, variable criteria | Consistent macro- and micro-categories |
| Sentiment | Subjective interpretation | Positive, negative or neutral, automated |
| Consistency | Varies depending on who reads it | Same criteria applied to every comment |
Server, surveys and export
When a tool analyses open-ended text with AI, it is reasonable to ask where that information travels. In the case of the AI Text Analysis Agent, processing happens locally, on RateNow's own servers.
It works with any survey you manage in RateNow — whether it is a patient experience, customer experience, employee experience or any other listening programme — and in multiple languages.
All comments can also be exported in Excel or CSV format. Simply go to "Create Report" and choose the "Comments XLS" or "Comments CSV" format. This allows you to analyse them, cross-reference them with other data or share them.
All within RateNow's servers
The AI Text Analysis Agent works locally. There is no transfer of information to third parties: the analysis remains within RateNow's infrastructure at all times.
Customer, patient or employee feedback is never in short supply — what is lacking is the time to process it at the pace it arrives. Reading comment by comment works at low volumes, but becomes unsustainable as the operation grows. Automating it does not replace the judgement of your CX team: it identifies the sentiment and category of each comment so that judgement can be applied to what truly requires a human decision.
It is already available in your RateNow account. If it does not appear in your dashboard, your account manager can activate it in moments.
How accurate is the AI Text Analysis Agent?
It achieves an accuracy above 85% in comment classification — a level comparable to the judgement of a human analyst.
What are the macro-categories and micro-categories of the AI Text Analysis Agent?
Macro-categories are broad thematic blocks (Emotion, Experience, Staff, Time, Resources). Within each one, micro-categories detail the specific reason — for example, within Emotion: gratitude, complaint or opinion.
Does the AI Text Analysis Agent work for any type of survey?
Yes, it works on patient experience, customer experience, employee experience surveys or any other feedback programme, in multiple languages.
Is there an additional cost for the AI Text Analysis Agent?
No. The feature is included in the RateNow service within the monthly subscription, with no additional configuration required.
Tell us about your project and get:
Best practices applied to your industry Surveys tailored to your needs Demo of our analytics platform A customized quotationⓘ No commitment
Tweet
Share
Share
