AI visibility

How ENERTRAG appears in ChatGPT, Claude and Perplexity, measured with a dedicated AI Visibility Score.

Partners, municipalities, investors and trade press increasingly research with AI answer engines. These systems rely mainly on third-party sources: trade media, Wikipedia, local press, YouTube. The company’s own website is one source among many. For ENERTRAG, we examined how the brand appears in these answers, where that picture comes from and which sources ENERTRAG can influence itself.

Starting point

Partners, municipalities and trade media increasingly ask AI systems about ENERTRAG. It was unclear how the brand is portrayed there, which sources shape that picture and what the team itself can change.

What we built

A dedicated AI Visibility Score across ChatGPT, Claude and Perplexity, a source map with six source types and template formats based on the Wensickendorf reference project.

What the team now does itself

Marketing, press and project communications follow a six-step roadmap, maintain the knowledge base and check every quarter with a fixed question set how ENERTRAG appears in AI answers.

Company knowledge and fact-checking: a dedicated AI Visibility Score

The basis is a monitoring system we developed for ENERTRAG: a fixed set of questions asked in ChatGPT, Claude and Perplexity at regular measurement dates. The questions range from unaided category questions and recommendation questions from a municipal perspective to direct brand questions, one of them in English.

Each answer is rated against a codebook: whether ENERTRAG is mentioned and in which position, whether the system recommends ENERTRAG, how well the portrayal fits the brand, how accurate the facts are and which sources are cited. This produces an AI Visibility Score for each system and measurement date. A second person codes borderline cases independently.

We also record which topics the systems associate with ENERTRAG, from wind energy and green hydrogen to municipal partnership. This shows which parts of the positioning come through in the answers and which sources can change the picture.

Brand story and brand voice: one project, many sources

Using the Wensickendorf hydrogen plant, we showed how one reference project becomes several citable sources: a structured technical article with an FAQ block, a press text, a YouTube film with chapter markers and corrected subtitles, infographics and a keyword library for all formats. A technical article states the key facts in the first paragraph, has a clear heading hierarchy and lists sources in the text and at the end. Film, article and press text contain the same key facts.

What the team now does itself: measure and refine

The roadmap has six steps. It starts with the audit, followed by a fact check and a maintained knowledge base. Then come writing rules per channel, machine-readable image and video content and quarterly measurement with the same question set and score. Each step has an owner at ENERTRAG and is designed so that the team can continue it independently.

This makes AI visibility a routine task for the communications department: measurable, reviewed quarterly and tied to the sources ENERTRAG can influence itself.

More Au·de case studies

Victorinox

Dräger

Hasso Plattner Institute