# Sentiment in the AI Era

> What AI sentiment is: how ChatGPT and other models describe a brand, why it differs from traditional sentiment, and how to add it to reputation monitoring.

URL: https://geosnap.ai/en/blog/sentiment-nell%E2%80%99era-ai
Language: English
Version IT: https://geosnap.ai/blog/sentiment-nell%E2%80%99era-ai
Author: Rinald Sefa, CMO Geosnap · Category: GEO strategy
Publisher: Geosnap (Maind Group S.r.l.), https://geosnap.ai

## In short

AI sentiment is the way generative models like ChatGPT describe a brand: its perceived strengths and weaknesses and the contexts in which they recommend it. Unlike sentiment on social media and reviews, it is a consolidated synthesis of the information ecosystem over time, often out of step with current positioning, and it can only be improved by working on that ecosystem. It is measured by systematically querying the models and should sit alongside existing brand monitoring tools.

- AI sentiment has three dimensions: perceived strengths, perceived weaknesses, and the context of recommendations.
- It reflects the information ecosystem over time, so it can stay tied to past positioning or past problems.
- Measure it by systematically querying models and benchmarking competitors, alongside existing brand monitoring.

Translated from the Italian original. [Read the original](https://geosnap.ai/blog/sentiment-nell%E2%80%99era-ai)

## A familiar metric, a new context

Anyone who works in brand reputation is well acquainted with the concept of sentiment analysis. Measuring how a brand is perceived, whether mentions are positive, negative, or neutral, is an established practice in modern marketing. There are dedicated tools, proven methodologies, and reference benchmarks. It is a metric that many companies monitor regularly on social media, on review platforms, and in traditional media.

In the era of generative AI models, this metric takes on a new dimension, and almost no one is measuring it in the right context yet.

When a user asks ChatGPT to recommend a solution for a particular need, the model does not just cite the relevant brands. It describes them, puts them in context, highlights their perceived strengths and weaknesses, and recommends them or advises against them with specific reasons. What emerges from these descriptions is a precise sentiment, one that directly influences the potential customer's perception, often before they have even visited the brand's website or read a single review.

## What AI sentiment measures

AI sentiment is not simply positive, negative, or neutral. It is a much more nuanced dimension that reveals how generative models build the narrative around a brand.

The first dimension is perceived strengths, meaning the characteristics that AI models systematically associate with a brand when they describe it. These are not necessarily the characteristics the brand wants to communicate; they are the ones that emerge from the information ecosystem the models were trained on. A brand might discover that AI models associate it mainly with ease of use, when its actual positioning is centered on technological innovation. Or that it is perceived as a solution for small businesses, when its actual target is enterprise companies.

The second dimension is perceived weaknesses, meaning the shortcomings that AI models tend to associate with a brand in their descriptions. These can be real weaknesses, reflecting negative feedback that is well established in the information ecosystem, or distorted perceptions that stem from outdated information or from an unbalanced information presence on certain topics.

The third dimension concerns the context of recommendations, meaning for which use cases and for which type of user AI models tend to recommend a brand. A brand can have an overall positive sentiment but be systematically recommended for a market segment that is not its strategic priority.

## The misalignment between real sentiment and AI sentiment

One of the most interesting things that emerges when you start measuring AI sentiment is how often it diverges from the brand's own self-perception and from the perception that emerges from traditional channels.

A brand that has worked in recent years to build a reputation for technological innovation may discover that AI models still describe it with attributes tied to its previous, more traditional and conservative positioning. This happens because the information ecosystem the models were trained on reflects a longer history than the recent one, and changes in positioning take time to settle into the external sources that the models use as a reference.

A brand that has fixed a notoriously serious customer service problem may discover that AI models still cite that issue as a weakness, because the negative discussions of the past are more numerous and more deeply rooted in the information ecosystem than the more recent positive mentions.

These misalignments are not errors in the traditional sense. AI models are simply reflecting the information available. But they have concrete consequences for the perception that potential customers form, often at a stage of the buying journey when the brand has not yet had a chance to tell its side of the story.

## Why AI sentiment is different from traditional sentiment

Sentiment on

social media or on review platforms reflects opinions expressed directly by users. It is a sentiment that the brand can monitor in real time, respond to, and influence through direct interventions on the customer experience or on communication.

AI sentiment works differently. It is not the sum of users' opinions at a given moment; it is a consolidated synthesis of the information ecosystem over time, processed and returned by the models in a way that is not completely predictable. You cannot respond to a ChatGPT description the way you respond to a negative review on Google. You cannot request the removal of inaccurate information the way you can on Trustpilot.

What you can do is work on the information ecosystem that feeds the models, building over time a presence that is richer, more consistent, and more aligned with the brand's actual positioning. It is slower and less direct work than managing traditional sentiment, but it is the only approach that produces real effects on AI sentiment in the medium term.

## Why most brands do not know their AI sentiment

The reason AI sentiment remains invisible to most companies is the same reason AI visibility in general remains invisible. There are no traditional tools that measure it. GA4 does not track conversations with generative models. Social listening tools do not include ChatGPT or Perplexity responses within their scope of analysis.

The only way to understand a brand's AI sentiment is to systematically query the models, analyze the descriptions they produce, identify recurring patterns in the strengths and weaknesses they attribute, and compare your own brand's sentiment with that of competitors. It is work that requires method and continuity, and that most companies have not yet started doing.

The result is that many companies are operating with a significant blind spot in their brand reputation. They know how they are perceived on social media, on review platforms, and in the media. They do not know how they are described by AI models, which is the tool that a growing share of their potential customers use to form a first opinion before they even search for them on Google.

## How to integrate AI sentiment into brand reputation monitoring

For companies that are starting to work on this dimension, the most practical starting point is to add AI sentiment alongside existing brand monitoring tools, without replacing them.

AI sentiment does not tell the same story as sentiment on social media or in reviews. It tells something different and complementary: how generative models synthesize and convey a brand's reputation to people researching in the early stage of the buying journey. It is an additional dimension that completes the picture, not a replacement for what already exists.

In practical terms, this means defining a set of queries representative of your market, systematically analyzing the descriptions that the models produce about your brand and your competitors' brands, identifying recurring patterns in sentiment, and monitoring how they evolve over time. The data collected this way goes into brand reputation reporting as a metric in its own right, with its own trend and its own strategic implications.

## A metric that will become standard

AI sentiment today is where social listening was fifteen years ago: a tool that some pioneering companies are beginning to explore, while most have not yet grasped its strategic relevance.

The difference is that the context is evolving much faster. The adoption of generative AI models in the pre-purchase research phase is growing significantly, and with it the relevance of AI sentiment for anyone who needs to understand how their brand is perceived at the most critical moments of the buying journey.

Those who start measuring it now build a historical baseline that will become increasingly valuable as the topic enters the mainstream brand reputation agenda. And they gain the ability to read a dimension of brand perception that competitors are probably not looking at yet.

[Watch the video on YouTube](https://www.youtube.com/watch?v=BE-dMS9s1H4)

## Frequently asked questions

### What is a brand's AI sentiment?

It is the way generative AI models describe a brand when they mention it: which strengths and weaknesses they attribute to it and for which use cases or types of user they recommend it. It shapes the potential customer's perception, often before they visit the website or read a review.

### Why does ChatGPT describe my brand differently from how it is positioned today?

Because the models reflect the information ecosystem they were trained on, which tells a longer story than the recent one. Changes in positioning or problems that have already been fixed take time to settle into the external sources the models use as a reference.

### How can I measure and improve my brand's AI sentiment?

Define a set of queries representative of your market, systematically analyze the descriptions models produce about your brand and your competitors, and monitor the patterns over time. To improve it, work on the information ecosystem by building a presence that is richer, more consistent, and aligned with your actual positioning.