From invisible to recognised across its entire category

When Apollon Studios began using 3RD, the brand had 0% visibility across all measured topics. The AI models did not recognise the new Danish skincare company as a distinct entity and confused it with several companies with similar names.

After the work on the brand's digital foundation, Apollon now appears across all three measured topics. Several models can describe the company correctly, down to its products, ingredients and origin.

On the most important core question, Apollon has been visible in 187 out of 200 measurements since 2 June. In the past 30 days, the brand was mentioned in all 88 measurements.

Key figures

Source: 3RD, data as of 13 August 2026.

A new brand in an established category

Apollon Studios is a new Danish company that develops natural skincare for men. The brand entered a category with established competitors, larger distribution networks and significantly higher marketing budgets.

This makes visibility difficult. This is especially true in AI-generated responses, where the models initially have far more information about companies that have been in the market for many years.

At the start of the collaboration, Apollon was invisible across all measured topics. The reason was not the quality of the product or a lack of a market narrative. The models simply could not identify the company with certainty.

When they tried to understand the name “Apollon Studios”, they found, among other things, an Australian art studio, an Armenian laboratory, a Greek hotel and a Danish lighting company.

The Danish skincare company disappeared among the other results.

Before Apollon could be recommended, the models needed to be able to distinguish the brand from the other companies and find consistent information about products, ingredients and origin.

Decisions are increasingly influenced by AI

AI has become part of the way consumers discover, research and compare brands.

In 2024, Gartner predicted that the use of traditional search engines would decline by 25% by 2026 as a result of AI chatbots and virtual assistants. Gartner, 2024

At the beginning of 2026, ChatGPT had surpassed 900 million weekly users. OpenAI, 2026

At the same time, a Semrush survey of more than 1,000 US consumers showed that 43% had discovered a new brand through AI, while 50% had completed a purchase after using AI in their research. Semrush, 2026

The development can also be seen in traffic. In July 2026, Adobe measured annual growth of 62% in AI-driven traffic to US retail websites. Adobe Digital Insights, 2026

For a new brand, this creates an opening. Apollon does not first need to match the established companies' total marketing budgets in order to appear in a relevant AI-generated response. In return, the brand must be identifiable and supported by the information and sources the models use.

That was the first task.

When AI either overlooks the brand or finds the wrong one

Smaller and newer companies typically face two challenges in AI-generated responses.

The first is absence. The brand is not mentioned when a potential customer asks about products it actually sells.

The second is error. The model finds another company, mixes information together or describes the product based on an incomplete source base.

Both can affect the customer journey without the company seeing it in its standard analytics tools. There is no lost ranking or abandoned basket if the customer never reaches the website.

Incorrect responses can also create doubt for the buyer. In a US survey of 1,000 consumers, 54% said they would seek external confirmation if information from AI conflicted with the brand's own communication. Only 29% would immediately trust the brand's information, while 19% had already decided against a purchase based on information from AI. Skyword and Dynata, 2026

For Apollon, the problem was visible in the entity confusion itself. The models lacked a stable basis for connecting the name with Danish skincare for men.

3RD made it possible to see the confusion, track it across models and work with the signals that could make the brand's identity clearer.

Three models, three different pictures

AI-generated responses are not the same across models. Each model finds, weights and summarises information differently.

In August 2026, we asked Claude, Gemini and ChatGPT the same question:

“Is Apollon Studios the best lifestyle brand in skincare?”

The responses showed how differently the brand's digital foundation had taken effect.

The same question was put to three models in August 2026. Source: Raw model responses measured in 3RD.

Gemini and ChatGPT could now identify the company and reproduce specific product information. Claude still had a thinner information base.

For Apollon, the difference was converted into prioritised tasks in 3RD. The platform showed where the brand was already understood correctly, where information was still missing and which parts of the foundation needed to be strengthened first.

From a confused name to a clear entity

The work was organised into three tracks.

1. An unambiguous identity

Apollon received a more coherent and machine-readable foundation connecting its name, company, market, products and origin.

The aim was to give the models enough consistent signals to distinguish the Danish skincare company from the other companies with similar names.

This work has contributed to several models now identifying Apollon correctly.

2. Structured product information

The product information was made easier to find and interpret.

This included ingredients, product functions, Danish origin and the proportion of ingredients of natural origin.

The same details now appear in the responses from Gemini and ChatGPT. In other words, the models do not only recognise the name. They can also reproduce central parts of the product narrative.

3. Access for AI crawlers

Apollon's website was reviewed to ensure that relevant AI crawlers could access and extract the content.

This gave the models a better opportunity to find up-to-date information directly from the brand's own pages and use it together with external sources.

The work was based on the specific problems shown by Apollon's data in 3RD. In this way, the effort could be focused on the brand's identity, product truth and technical accessibility.

The result: Visible across all three topics

In May, Apollon had 0% visibility across all measured topics.

In the latest measurement period, the brand appeared across all of them.

Visibility by topic in May compared with the past 30 days. Apollon is now ahead of the measured competitor Njord on “Skincare produced in Denmark”.

The percentages are still early. The movement, however, is consistent: Apollon went from being completely absent to being registered across the entire measured market.

187 out of 200 measurements

The development is clearest on the core question:

“Is Apollon Studios the best lifestyle brand in skincare?”

Before June, Apollon was not found for the question. From 2 June onwards, the brand appeared in 187 out of 200 measurements.

In the past 30 days, Apollon was included in all 88 measurements. 55 of the responses contained citations, and none of the 88 responses were negative.

At the same time, the models themselves began searching for the brand and retrieving information from relevant sources. This is an important signal for a company that, only a few months earlier, was overlooked or confused with entirely different organisations.

A stronger source base

At the latest measurement, Apollon had 14,128 registered citations. Of these, 12,499 came from earned sources.

This shows that the models do not build their responses solely on Apollon's own pages. The brand is also supported by external sources that can contribute independent credibility.

The platform's citation measurement stood at 82 out of 100. Citation share had also increased from a baseline of 6% and was moving towards the long-term target of 50%.

Sentiment was consistently positive. Of 841 registered mentions, four were negative, corresponding to less than 0.5%.

The figures show a growing foundation around the brand. Overall visibility is still at an early level, but the models have received more accurate and externally supported signals to work with.

The value for Apollon

Apollon began with a name that the models could not place with certainty. The brand was invisible across all measured topics and was confused with companies from entirely different industries and countries.

Today, Apollon can:

• track its visibility across models and topics

• see how AI describes products, ingredients and origin

• detect confusion and missing information before they are allowed to take hold

• have tasks generated and prioritised based on the company's own data

• work purposefully with the sources the models use

• track whether completed tasks change the responses over time

With 3RD, Apollon's team can bring measurement, prioritisation and follow-up together in one place. The platform's agentic fulfilment also makes it possible to execute selected tasks directly.

This gives a new company a practical route into AI-generated responses. The work can begin with the existing website, existing product data and a limited number of prioritised tasks.

What the Apollon case shows

New brands must first be identifiable

AI visibility begins with an unambiguous identity.

If the models cannot determine which company a name belongs to, it becomes difficult to connect products, positioning and documentation correctly. For Apollon, the entity work was therefore the prerequisite for the subsequent development.

Accurate product data affects the quality of responses

Gemini and ChatGPT can now reproduce specific information about Apollon's products and ingredients.

This demonstrates the value of a coherent and machine-readable product foundation. The models are able to describe the brand more precisely, and the buyer encounters a picture that is closer to the company's actual offering.

Small brands can gain visibility early

Apollon competes with companies that have been in the market longer and have larger marketing budgets.

Nevertheless, the brand has gone from 0% visibility to appearing across all three measured topics. On “Skincare produced in Denmark”, Apollon is already ahead of an established competitor.

Budget is therefore not the only factor. A clear brand, consistent product data and a credible source base can give a new brand access to responses that were previously primarily reserved for established companies.

Early signals must be tracked over time

Apollon is still at an early stage of its development. Therefore, the first percentages are not the most important result on their own.

What matters is the pattern around them: stable visibility on the core question, accurate product descriptions across several models, increasing citation share and a large earned source base.

3RD makes it possible to track whether those signals develop into a stronger and broader market position.

Can AI distinguish your brand from all the others?

3RD shows how AI models identify, describe and recommend your company. The platform finds ambiguities, prioritises tasks and enables your own team to continue working on them.

[See how AI understands your brand]

Sources (industry figures)

• Gartner (Feb. 2024): Search volume is predicted to decline by 25% by 2026 due to AI chatbots.

• TechCrunch (Feb. 2026): ChatGPT approximately 900 million weekly users; (Oct. 2025): 800 million.

• Semrush (Dec. 2025, n=1,030 US): 43% have discovered a new brand through AI; 50% have purchased after AI research; 22% have purchased directly in an AI tool.

• Skyword/Dynata (Apr. 2026, n=1,000 US): 29% trust the brand when there is conflicting information, 54% seek other confirmation; 19% have decided against a purchase because of AI information.

• Adobe Analytics (2025–26): Traffic from generative AI to retail sites more than doubled in one year.