Sentiment Ratings Dashboard

Monitor satisfaction in services via tato's AI analysis.

The Sentiment Ratings Dashboard uses AI to analyze the tone and quality of service interactions, classifying them into five sentiment categories (Terrible, Bad, Fair, Good, Excellent).

What does tato's AI observe? Communication, problem resolution, and overall interaction quality, providing objective insights on service performance that can be compared with customer ratings.

How to Access

Access: Service / Digital Service / Dashboards / Sentiment Ratings, as shown in Figure 1:

^{Figure 1: Sentiment Ratings Dashboard.}

---

What is AI Sentiment Analysis?

tato's Artificial Intelligence analyzes each completed service interaction, generating an evaluation based on multiple criteria:

• Quality of communication between agent and customer;

• Effectiveness in resolving the presented problem;

• Clarity and depth of the responses provided;

• Tone and approach used during the interaction.

Each interaction receives an explanatory comment and a sentiment rating, allowing an objective and consistent view of operational quality without relying exclusively on customer feedback.

---

Available Actions on Charts

Each dashboard chart has a set of actions that allow you to customize the view and export the data. These controls appear in the upper right corner of each chart:

^{Figure 2: Available Actions on the Charts.}

1. Hide Chart

The first icon allows you to temporarily hide the chart from the dashboard view. Useful when you want to focus on other specific indicators without losing the chart configuration.

Click the icon to view the chart again.

2. Maximize Chart

The second icon expands the chart to full screen, allowing more detailed analysis of the data and better visualization of complex information.

3. Filters

The third icon opens the filter panel specific to each chart. Available filters vary according to the chart type (detailed in the following section).

4. Export

The fourth icon allows you to export the chart data in the formats:

• SVG - vector format ideal for presentations and documents

• PNG - image format for quick sharing

• CSV - spreadsheet format for additional analysis in Excel or other tools

---

Charts and Their Filters

Below, learn about each chart available on the dashboard and its specific filters:

1. Sentiments by Day

Displays the distribution of sentiments identified by the AI throughout the day, segmented by category (Terrible, Bad, Fair, Good, Excellent), showing the absolute number of interactions in each classification.

^{Figure 3: Sentiments of the Day.}

The chart uses colored bars representing:

• Red = Terrible

• Orange = Bad

• Yellow = Fair

• Light green = Good

• Dark green = Excellent

Available filters:

• Period - select the start and end date for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

2. Daily Sentiment Success Rate

Shows the daily evolution of the proportion between interactions with positive sentiments (green) and the total analyzed interactions (orange), indicating the consistency of quality over the days.

^{Figure 4: Daily Service Success Rate.}

The chart presents stacked bars where:

• Green = Interactions with Positive Sentiments

• Orange = Interactions with Sentiments (total)

Available filters:

• Period - select the start and end date for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

3. Sentiments by Month

Presents the monthly distribution of sentiments identified by the AI, segmented by category, allowing identification of quality trends over the months.

^{Figure 5: Sentiments of the Month.}

The chart shows the absolute number of each sentiment category (Terrible, Bad, Fair, Good, Excellent) from a monthly perspective.

Available filters:

• Month/Year - select the month and year for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

4. Monthly Sentiment Success Rate

Displays the monthly evolution of the positive sentiment rate through stacked bars, facilitating performance comparison between months.

^{Figure 6: Monthly Sentiment Success Rate.}

The chart presents monthly stacked bars where:

• Green = Interactions with Positive Sentiments

• Orange = Interactions with Sentiments (total)

Available filters:

• Month/Year - select the month and year for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

5. Sentiments by Year

Shows the consolidated annual distribution of sentiments identified by the AI, segmented by category, offering a macro view of service quality throughout the year.

^{Figure 7: Sentiments by Year.}

The chart displays the count and percentage of each sentiment category (Terrible, Bad, Fair, Good, Excellent) from an annual perspective.

Available filters:

• Year - select the specific year for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

6. Annual Sentiment Success Rate

Presents the evolution of the positive sentiment rate across the months of the year, showing the consistency of service quality from an annual perspective.

^{Figure 8: Annual Sentiment Success Rate.}

The chart shows monthly stacked bars where:

• Green = Interactions with Positive Sentiments

• Orange = Interactions with Sentiments (total)

Available filters:

• Year - select the specific year for analysis

• Department - filter specific departments

• Users - filter by specific agents

---

7. Sentiments by Service Group

Compares the distribution of sentiments among different service groups, revealing which types of requests generate better or worse perception according to the AI analysis.

^{Figure 9: Sentiments by Service Group.}

The chart uses stacked bars showing the percentage proportion of each sentiment category by service group.

Available filters:

• Period - select the start and end date for analysis

• Service Group - filter specific groups

---

8. Sentiments by Department

Compares the distribution of sentiments among different departments, allowing identification of areas with better or worse service quality according to the AI analysis.

^{Figure 10: Sentiments by Department.}

The chart presents stacked bars showing the percentage proportion of each sentiment category by department.

Available filters:

• Period - select the start and end date for analysis

• Department - filter specific departments

---

Interpretation of Sentiments

AI Rating Scale

The Artificial Intelligence classifies interactions into five categories:

• Excellent - Exemplary service with clear communication, effective resolution, and a professional approach

• Good - Satisfactory service that meets expected standards

• Fair - Average service with identified areas for improvement

• Bad - Service with significant problems in communication or resolution

• Terrible - Service with serious failures that compromise quality

Difference between Customer Rating and AI Sentiment

It is important to understand the differences between these two metrics:

Aspect	Customer Rating	AI Sentiment
Origem	Direct feedback from the customer	Automatic analysis by the AI
Coverage	Only customers who respond	All interactions
Objectivity	Subjective (personal perception)	Objective (technical criteria)
Timing	After closure	Continuous during analysis
Focus	Overall satisfaction	Technical quality of the service

Using both metrics together offers a 360° view of quality: the customer's perception (emotional) and the objective technical assessment (AI).

Performance Indicators

• Above 70% of "Good" and "Excellent" sentiments = Excellent performance

• Between 50% and 70% positive sentiments = Adequate performance

• Below 50% positive sentiments = Requires immediate attention and corrective actions

---

Use Cases

The Sentiment Ratings Dashboard is especially useful for:

• Objective and consistent evaluation - measuring quality with standardized criteria by the AI in 100% of interactions, not only in those that receive manual ratings.

• Identification of training gaps - detecting agents or departments with low sentiment ratings to target specific training.

• Cross-validation with manual ratings - comparing customer perception with the AI's objective analysis to identify discrepancies and opportunities.

• Real-time quality monitoring - tracking sentiments without relying on customer engagement in satisfaction surveys.

• Performance analysis by request type - identifying which service groups present greater quality challenges according to technical criteria.

• Benchmarking between teams - comparing departments fairly using objective metrics generated automatically.

• Automated quality audit - generating documented evidence of the technical quality of interactions for certifications and audits.

• Detection of low-quality patterns - quickly identifying deterioration in service quality before it impacts customer satisfaction.

• Continuous feedback for agents - using AI comments as a basis for development conversations and individual improvement.

• Prioritization of operational improvements - directing resources to areas with a higher concentration of negative sentiments according to objective analysis.