Build Guide: Setting Up AI Reporting
- Gnatta
- Abbey Brown
Quick links in this article:
Once your AI flows are up and running — capturing Intent, Sub-Topic, and Entity data — you can use Gnatta’s new reporting tiles to visualise trends and spot patterns across your customer conversations.
This guide walks you through creating Tree Map and Correlation tiles that bring your AI-powered data to life.
Before You Start
Before you can build your reports, you’ll need to make sure you’re feeding data into Gnatta in the right format. Try following the guides below first:
✅ Build Guide: Intent and Sub-Topic Routing
Make sure both fields are stored in Single Select dropdowns. You don’t have to complete the routing steps or update queues — but you do need to collect the data cleanly.
✅ Build Guide: Entity Recognition for Reporting
Ensure you’re detecting entities like products, organisations, or locations and storing them in open text custom fields.
✅ Build Guide: Sentiment-Driven Routing
Make sure Sentiment is being stored in a Single Select dropdown. You don’t have to route with this data, but you do need to collect the data!
Once you’ve captured some data, you’re ready to start visualising it.
Create or Choose a Report
Go to Reporting.
Select the report you want to add your tiles to — or create a new one from scratch: Reporting Suite
You can mix AI tiles with your usual performance and volume metrics. It’s all in one place.
About These Tiles
Gnatta’s AI-powered reporting tiles help you explore trends in Intent, Sub-Topic, and Entity data.
Here’s a quick overview of the four tile types:
Tile | Best For |
|---|---|
Overview Tree Map | Layered breakdowns – e.g. Reason → Sub-Topic → Product |
Top 100 Tree Map | Deep dives into one Reason – shows the top 100 values for an Entity field |
Top 10 Tree Map | Comparing multiple Reasons – shows top 10 Entity values for each |
Correlation Radar | Spotting relationships between dropdown fields – e.g. Sentiment by Reason |
Overview Tree Map Tile
This tile gives you a layered breakdown of AI-detected data — for example:
Reasons split by Sub-Topic and then Products.
Click + Add Tile and choose Closed by Data Treemap – Overview
Set the tile Name (e.g. Reasons by Product)
Choose the following data fields:
Reason (must be a dropdown field type, select one or all values)
Sub-Topic (must be a dropdown)
One AI Entity field (e.g. Product Mentioned or Organisation Mentioned)
Once you’re happy with the config, click Save.
Top 100 Tree Map Tile
This tile gives you a focused view of volume by entity, within a specific reason for contact — for example: Which products are most often mentioned in Refund requests?
It’s ideal for spotting trends and problem areas tied to a particular Intent, and can display the top 100 values all in one tile.
To build the tile:
Click + Add Tile and choose Closed by Data Treemap – Top 100
Set the tile name (e.g. Top Refunded Products)
Choose:
Reason – e.g. Return and Refund (must be from a dropdown field)
Entity Field – e.g. Product Mentioned
Use this tile to monitor things like:
Most-refunded products
Website issues tied to specific organisations
Most-mentioned locations in delivery complaints
Once you’ve configured it, click Save to add it to your dashboard.
Top 10 Tree Map Tile
This tile works just like the Top 100, but gives you a broader view across multiple Reasons, showing the Top 10 results within each.
For example: See the top 10 most-mentioned products in Refunds, Tracking, and Website Issues — all in one tile.
To build the tile:
Click + Add Tile and choose Closed by Data Treemap – Top 10
Set the tile name (e.g. Top 10 Products by Reason)
Choose:
Reason – select multiple values (e.g. Return and Refund, Website Issues, Parcel Tracking)
(Must be a dropdown field)Entity Field – e.g. Product Mentioned (must be open text)
This tile is designed for comparison - it’s great for surfacing emerging trends across different contact types, even if it doesn't go as deep as Top 100.
Click Save to finish.
Correlation Radar Tile
The Correlation Radar helps you identify relationships between two dropdown fields — like how often certain Reasons are linked to a specific Sentiment.
This tile is best used when:
Both data fields are Single Select dropdowns
Each field has a limited number of values (e.g. 3–6)
There’s likely to be a meaningful pattern
To build the tile:
Click + Add Tile and choose Closed by Data Correlation
Set the tile name (e.g. Sentiment by Reason)
Choose:
Reason – must be a dropdown field (e.g. Intent or Topic)
Sentiment – must also be a dropdown (e.g. using AI sentiment detection)
Example use case: Spot which Reasons are most often flagged as Negative or Mixed sentiment to help with targeting escalations or pain points.
Once your fields are selected, click Save. You’ll see each Reason plotted with its associated sentiment breakdown — giving you a quick, visual sense of customer mood by topic.
What’s Next?
Now that your AI dashboards are live, keep an eye on the patterns they reveal:
Which products are driving the most contact?
Which reasons are trending negative?
Where can automation or self-serve reduce volume?
As your AI flows collect more data, these tiles will only become more valuable — helping you prioritise fixes, improve customer journeys, and drive smarter decisions across the business.