AI Agents Just Got Smarter: 3 New Features to Automate Feedback Analysis with More Context and Precision

In June, Feedier reached a new milestone in intelligent feedback automation. Three new AI agents have been deployed to boost the efficiency of your operations, bringing more context, greater accuracy, and improved detection of weak signals. With these new features, your teams save time, benefit from proactive recommendations, and gain more reliable insights generated directly by the platform.

An AI Agent to Automatically Describe Your Teams

Feedier has introduced a new Admin Agent that can automatically generate a personalized description for each team. These descriptions provide valuable context to other AI agents, helping them better understand each team's scope and responsibilities, so they can tailor their recommendations accordingly.

Why does it matter?
Each team has a specific role in your organization. By adding a description, you give the AI agents the context they need to sharpen their analysis and deliver more relevant suggestions.

How does it work?
The agent runs automatically every Monday. It identifies teams that don’t yet have a saved description and generates a short, contextualized one based on the team name and available organizational data. The suggestion appears in the Recommendation Center, and a notification is sent to users with the appropriate agent access.

2 options :

1. If you accept the suggestion, the description is saved and the agent stops sending reminders for that team.
2. If you reject it, the agent will try again the following week, until a description is manually added or the agent is disabled.

An AI Agent to Refine Topic Instructions Based on Real Usage

Feedier now makes it easier to manage and fine-tune topic instructions with a new AI agent built into the text analysis module. This agent, enabled by default, learns from your manual tagging activity, especially when you add or remove responses from a topic, to detect cases where the original instruction no longer matches the reality of your data.

Each week, the agent reviews the last seven days of tagging behavior. If inconsistencies are detected, it suggests a new and improved version of the instruction. Users receive a notification and can view the proposal in the Recommendation Center. Accepting it instantly updates the topic with the new instruction, while rejecting it keeps the current one in place. This feature delivers value on two levels: it reduces the need to manually review all topic instructions, and it improves the accuracy of future automatic tagging. In short, it’s a direct and intelligent way to continuously improve your text analysis, driven by how you actually use the platform.

An AI Agent to Automatically Detect Anomalies in Your Attributes

Feedier’s newest AI enhancement, the Analyst Agent, helps you spot unusual patterns in your feedback data, especially when it comes to the Satisfaction Ratio of your most important attributes. Whether performance is significantly better or worse than expected, the agent flags these anomalies and guides you to take meaningful action.

The agent runs automatically every Monday. It reviews the past five weeks of attribute data, focusing on values that are statistically relevant, those that represent at least 1% of your weekly feedback and have collected five or more responses. Using a z-score threshold of ±2.5, it identifies outliers based on both historical trends and comparisons with other attribute values.

When an anomaly is detected, Feedier generates a ready-to-use report tailored to the situation. No setup is needed: the filters, KPI, time range, and comparison logic are already pre-filled.

Here’s what you get and can do:

• Clear anomaly labels: “Positive anomaly” (value performs above average) or “Negative anomaly” (performance drop)
• Pre-filled custom report: Last 7 days, Satisfaction Ratio KPI, anomalous value vs. others
• Instant decision-making: Accept, reject once, or stop tracking the attribute
• All suggestions stored in the centralized Recommendation Center
• Smart notifications sent to users with the right access level

This new agent gives you the tools to catch weak signals early, react faster, and stay in control, without spending time manually scanning for trends.

Conclusion

With these three new intelligent agents, we’re reinforcing our core promise: turning feedback analysis into a smooth, intelligent, and action-oriented process. By automatically enriching organizational context, refining instructions based on real usage, and proactively detecting anomalies, the platform helps you save valuable time while improving the quality of your decision-making.

This intelligent automation doesn’t replace humans, it enhances them. Every recommendation is contextualized, transparent, and always under your control. You remain in charge, but you’re no longer alone in navigating the growing complexity of your data.

These AI agents embody a new way of managing feedback: proactive, personalized, and continuous. It’s another step toward a more controlled customer experience, and an organization that learns and adapts constantly.

‍