AI glossary · Business
Sentiment analysis
In one sentence
Sentiment analysis is the AI technique that detects the emotional tone of a text (positive, negative or neutral, and sometimes emotions such as anger or urgency) so you know how the writer feels without reading everything by hand.
Also known as: opinion mining · sentiment detection · emotion AI · tone analysis
What is sentiment analysis?
It’s a form of text classification where the question isn’t “what is this about?” but “how is the writer feeling?”. It’s also called opinion mining, because it started out as a way to process reviews and social media comments at scale.
Put it to work
An agent that sorts and answers your email
From €39/mo + VAT
It sorts the inbox by priority and drafts replies with context.
See the email agentThe output can be more or less detailed:
| Level | Output | Example |
|---|---|---|
| Polarity | Positive, negative, neutral | ”Negative” |
| Score | A numerical scale | −0.8 (very negative) |
| Emotion | Anger, frustration, joy… | ”Frustration” |
| Aspect-based | Tone for each topic in the message | Product: positive · Delivery: negative |
Aspect-based analysis is the most useful in customer service, because a single message often mixes good and bad.
Related to this term
UTIO Email
- Add-on for any voice or WhatsApp plan
- 2 mailboxes and 2,000 emails a month
- Standalone email: from €89/mo
From €39/mo + VAT
See the email agentHow it works
- The text comes in: an email, a message, a review or a call transcript.
- It’s processed with NLP: cleaned up and turned into a representation the model understands.
- The model assigns the sentiment. Older systems counted positive and negative words from a dictionary; language models understand context, negation (“I’m not happy at all”) and some irony.
- The result is put to use: raising priority, alerting a person, feeding a report.
The score is only as good as the categories behind it, so it’s worth checking a sample by hand every few weeks, especially messages in other languages or full of slang.
A very common use is as a trigger for human handoff: if a customer is clearly upset, a person should deal with them, not a bot.
Example
An online furniture shop in Exeter receives dozens of emails every Monday after the weekend deliveries. With sentiment analysis, messages with a very negative tone (a damaged item, a failed assembly) are flagged as priority and go straight to the customer service lead first thing. Thank-you notes are grouped separately and used to ask for a review.
Sentiment analysis at UTIO
UTIO’s email agent includes sentiment and urgency detection: it reads your mail, sorts it into threads and prioritises it, urgent at the top. That way a complaint doesn’t get buried under automated emails. Every message lands on the contact’s CRM record with an owner and a next step.
The agent drafts the reply in your tone, and you decide whether to review it before it goes out. For sensitive messages, keeping that review step is the sensible choice.
Frequently asked questions
How reliable is sentiment analysis?
It does well on clear messages but struggles with irony, sarcasm and mixed messages ('great product, terrible delivery'). Today's language models beat the old keyword-based systems, but it's best used to prioritise, not to make decisions on its own.
What's it useful for in a small business?
Mainly for not losing an unhappy customer: spotting complaints in email or WhatsApp and dealing with them first. It also helps track how reviews and survey answers are trending without reading each one.
What's the difference between sentiment analysis and emotion detection?
Sentiment analysis classifies polarity: positive, negative or neutral. Emotion detection goes further and tells anger, frustration, joy or worry apart. Many tools combine both with an urgency score.
Related terms
- Text classification
Text classification is the AI task of automatically assigning one or more categories to a piece of text, such as an email or a message: type of request, topic, priority or language. It's the basis for sorting, prioritising and routing work.
- Natural language processing (NLP)
Natural language processing (NLP) is the branch of artificial intelligence that lets machines work with human language: read, analyse, classify, translate and produce text or speech the way people use it.
- Human handoff
Human handoff is the point at which a chatbot or AI agent passes the conversation to a member of staff, together with the context it has gathered, because the case needs it, a rule requires it or the customer asks for it.
- Lead scoring
Lead scoring is a way of giving each sales contact a score based on how close they are to buying, combining who they are (sector, size, location) with what they've done (questions, visits, replies), so the sales team deals with the most promising first.
Updated on