By Daniel Espejo, Founder & CEO at Omnia. 1st of September 2025

How to Increase your Visibility in AI: The New Share of Voice Explained With Real Data

In this article

• The new reality: when brands disappear from AI responses

• Share of Voice has also been transformed

• Why measuring Share of Voice is no longer optional

• It's not just about mentions: understanding what feeds the models

• Real case study: From Sun With Love and its data-driven strategy

• What you can do today if you want to be visible tomorrow

• Conclusion

The new reality: when brands disappear from AI responses

Most brands are disappearing and they don’t even know it. Engines such as ChatGPT, Gemini, and Perplexity do not display links: they generate responses. And in those responses, only a few brands have a voice. The rest have disappeared.

Until now, we talked about Share of Voice as the percentage of impressions, clicks, or mentions that a brand had compared to its competitors. But that was before.

Today, we are discussing Share of Voice in generative environments, which is divided into two fundamental metrics:

• Domain Share of Voice: the percentage of times your domain is used as a source in responses generated by AI models. It indicates whether your content is considered reliable and valuable.

• Brand Share of Voice: the percentage of times your brand is mentioned in those same responses. It reflects whether your name is present in the decisions made by the model.

Both metrics are important, but they capture different aspects of your visibility. You can have Brand Share of Voice without Domain Share of Voice if models mention your brand but cite other sites as sources. And you can have Domain Share of Voice without Brand Share of Voice if your website is used for its answer but the model does not explicitly mention your brand.

Example: A media outlet may use your site as a source without mentioning your brand. Or your brand may appear as a recommended option, but the source that validates that recommendation will be another domain.

If you don't know these two metrics, you're blindly managing your presence on the fastest-growing channels. And if you don't know what sources are feeding those responses, you can't do anything to change it.

Omnia gives you just that: total visibility into how, where, and when your brand (and your website) appears in AI-generated responses. And what is happening with your competition.

In this article, we explain why this is critical and how an emerging brand, From Sun With Love, went from not existing in AI to surpassing Columbia, Decathlon, and Patagonia in mentions.

Share of Voice has also been transformed

Historically, Share of Voice (SoV) measured what percentage of visibility a brand had compared to its competitors. In traditional advertising, it was a question of media investment. In SEO, it was about Google rankings. In social media, it was about mentions or interactions.

However, the rise of generative models has forced us to redefine this metric. Visibility today depends not just on where you publish or what you say, but on what the models read, trust, and use.

We now talk about Brand Share of Voice and Domain Share of Voice. Both are important, but neither are useful if you do not understand where those responses are coming from — and how to influence them.

Why measuring Share of Voice is no longer optional

Knowing how many times your brand appears in AI responses is just the starting point. What really matters is knowing what is influencing those responses. It is not enough to appear once: you need to understand the dynamics that keep you in or out of the responses.

Some reasons why Share of Voice in AI belongs on your dashboard:

• The responses are limited: typically between 3 and 5 brands, compared to 10-20 links from Google.

• Models tend to solidify their responses: if a brand is already well represented, it is more likely to continue appearing.

• The user does not compare, they decide: if your brand appears, you have a real chance of converting. If not, you lose the opportunity without even knowing it.

Measuring SoV gives you visibility, but also urgency: if you do nothing, someone else will take that place.

It's not just about mentions: understanding what feeds the models

A common mistake is to think that having your own well-positioned content is enough. Many brands write hundreds of articles and wonder, ‘Why don't I appear in the results?’ 

But language models don't just consume content from your website. In fact, in many cases they prefer:

• Established media sources.

• Forums such as Reddit or Quora.

• YouTube

• Reports or educational resources.

If you do not understand which sources the model is reading, you cannot influence its response.

With Omnia, you can identify:

• The most frequently cited domains by sector and type of prompt.

• The competitiveness and traffic each prompt has.

• Where are your competitors publishing?

• Which channels do you need to be on to have an influence?

Real case study: From Sun With Love and its data-driven strategy

From Sun With Love is an emerging sun-protection clothing brand. When they began analysing their presence on AI engines with Omnia, the reality was that it did not appear in any relevant responses within their category.

Not because it was irrelevant, but because they were not present in the sources that the models considered reliable.

What Omnia showed:

• High-value prompts that triggered purchasing decisions.

• The brands that dominated those responses.

• The most frequently cited domains as sources: unotv.com, skincancer.org, reddit.com, youtube.com, among many others.

The strategy:

• They identified the media, channels, and platforms where they needed to appear.

• They collaborated with sources that were already feeding the model.

• They generated relevant, quotable content aligned with the detected prompts.

The result?

In less than 8 weeks, its Domain Share of Voice increased to 4.7%, competing with high-volume media sources.

But matters most: Their Brand Share of Voice reached 12.1%, overtaking giants such as Columbia, Decathlon, and Patagonia.

Today, they appear in AI-generated responses when a user searches for products in this category.

This case proves that visibility is not a matter of budget, but of information. If you know what to look for, you can take action.

What you can do today if you want to be visible tomorrow

Waiting for AI to magically recognise you does not work. There are specific actions you can take today:

1. Audit your current visibility

Do you appear in your sector’s key prompts?

Which brands are dominating the responses?

Which models (ChatGPT, Gemini, Perplexity) recognise you?

2. Analyse the sources cited.

How much does your domain matter?

What other sites are influencing the answers?

Where does your competition publish, and where are you missing?

3. Adapt your content strategy

Create specific content for decision prompts.

Publish in media outlets that are already being cited.

Optimise your content so that it is quotable, concise, and easy to summarise.

4. Keep track of progress

Measure your progress week by week with Omnia.

Adjust your strategy based on the results.

Conclusion

Generative SEO is not an extension of traditional SEO. It is a whole new field, with its own rules and sources of influence.

Visibility is not only for big brands with big budgets, every brand can be visible with the right information and the right strategy.

If you want to be in AI responses, you need to understand which prompts matter, who dominates them, what sources feed the system, and what your actual position is.

Request a demo and see how Omnia can show you where you are, where you should be, and how to get there.