The short version
HubSpot does not score sentiment on support tickets out of the box, and there is no native “customer mood” property waiting for you in a fresh portal.
Sentiment scoring on tickets has to come from somewhere: a human reading every ticket, a rule you write yourself, or an app that reads the ticket text and writes a score back.
This guide covers what sentiment scoring on tickets actually looks like in HubSpot, why it should land as a native ticket property rather than in a separate dashboard, and how to act on it once it exists.
Why sentiment on the ticket, not sentiment in a report you check later
Support teams already triage by pipeline stage, priority, and owner. Sentiment is most useful when it sits next to those signals on the same object, not in a tool you have to open separately after the fact.
If sentiment lives as a property on the ticket record, it inherits everything HubSpot already does with properties: it shows up in the ticket index, it can be a column in a view, it can be a filter, it can drive a workflow, and it can be a dimension in a custom report.
A standalone sentiment dashboard can look impressive in a demo, but it adds a second place to check, a second login, and a signal that is disconnected from the ticket your rep is actually working on.
What “native ticket property” means in practice
HubSpot lets you create custom properties on tickets from Settings, choosing a data type such as single-line text, number, or a dropdown (enumeration) with a fixed set of options.
Once a property exists on the ticket object, it behaves exactly like any built-in field: it appears when you edit a ticket, it can be added as a column, and it is available to workflows and to the custom report builder, with no extra integration work.
This is the mechanism an AI scoring app should use. It should add its scores as ticket properties, not as a note buried in the ticket timeline or a number that only exists inside its own interface.
How an AI scoring app actually does this
Here is the honest version of what happens under the hood, not the marketing version.
The app reads the ticket’s subject and description, the same text your support rep would read, and asks a language model to classify it.
It then writes the result back onto the ticket as a small number of custom properties. In SentimentSync’s case, that is four: sentiment, category, intent, and priority.
It does not read call recordings, attachments, or the full email thread history unless that content is already part of the ticket description. A one-line ticket with no real description gives the model very little to work with, and the score reflects that.
Scoring runs when a ticket is created or updated, and existing open tickets can be backfilled once the app is installed, but sentiment on a ticket that was closed months ago will not update itself.
Acting on sentiment once it is a property
With sentiment sitting on the ticket, you can build a ticket-based workflow in Automation that enrolls when the sentiment property is set or changes, and then take an action: notify the ticket owner, create a task, or change the ticket’s priority.
You can also add the sentiment property as a filter in your ticket views, so a support lead can pull up every open ticket scored as negative without reading the whole queue first.
For reporting, the property is available like any other field in HubSpot’s report builder, so you can track sentiment distribution over time or break it down by pipeline, owner, or source.
Setting expectations
Sentiment scoring is a signal, not a verdict. It is meant to help a team prioritize attention, not to replace a human reading a ticket before responding.
It works best on tickets that contain real text. A ticket that just says “see attached” will not produce a meaningful score, because there is nothing in the subject or description to read.
If you are evaluating this for your own support queue, start with a free tier that covers a real month of tickets before you commit to anything, and check the scores against your own read of a sample of tickets.