Brand Mentions in ChatGPT in the Financial Sector: Zeno Visibility Structures LLM Brand Monitoring Analysis for a Regulated Institution
Brand Mentions in ChatGPT in the…
Starting Point
A mid-sized, nationally operating financial institution with approximately 1,400 employees and a regulated product portfolio wanted to gain reliable insight in 2024 into how its brand appears in generative AI systems. The focus was on brand mentions in ChatGPT within the financial context — specifically whether the institution was being referenced at all when users asked typical questions about accounts, financing, investments, and trustworthiness. The team managed marketing, SEO, and content centrally, but operated in close coordination with compliance and specialist departments.
The starting position was ambivalent: classical SEO metrics were solid, but generative systems painted a different picture. In an initial LLM brand monitoring analysis, the brand appeared in only 6 out of 30 standardized prompts in ChatGPT. In Perplexity, it appeared in 7 out of 30, and in Claude just 4 out of 30. At the same time, competitors and generic advisory websites were being cited more frequently than the institution itself. While the existing content inventory included more than 900 indexable URLs, it lacked consistent semantic interlinking, a unified entity structure, and only partial Schema.org markup.
Challenge
The core problem was not the volume of content, but the absence of machine-readable authority. For generative models, the brand was insufficiently recognizable as a trustworthy source — particularly for high-trust queries around deposit protection, fees, product comparisons, regulatory classification, and suitability for specific target groups. In these contexts, LLMs are especially cautious and tend to draw on sources that are semantically clear, topically consistent, and structurally well-connected.
This had several consequences for the institution: First, brand perception within LLMs was not manageable. Second, there was no clean baseline to measure the actual impact of content initiatives. Third, the manual workload was high, as the marketing and SEO teams had to review responses across ChatGPT, Gemini, Perplexity, Claude, and Copilot separately. In a regulated environment, an additional layer of complexity arose: every new content component had to pass through specialist and legal review. Without structured monitoring, it remained unclear which topics deserved priority and which content actually improved LLM visibility.
Solution Approach
The institution chose Zeno Visibility because the platform not only delivers LLM brand monitoring, but directly translates the analysis into a semantic authority system. The approach was deliberately two-phased: measure first, then systematically close the gaps.
In Phase 1, a baseline was established across five LLMs using Zeno Visibility's research engine: ChatGPT, Gemini, Perplexity, Claude, and Copilot. The team defined 42 standardized prompts across three clusters: product intent, trust and compliance questions, and comparison and decision-making queries. This produced a Semantic Authority Score per topic area and model, making it visible where the brand was already present and where LLMs were instead defaulting to competitors or generic content.
In Phase 2, the findings were translated into the Authority System Builder. For 18 core keywords, Zeno Visibility generated a complete content system with more than 100 semantically interconnected assets per topic area — including hub pages, FAQ modules, comparison pages, case studies, and supporting specialist articles. The content was exported CMS-ready for the existing WordPress setup. In addition, JSON-LD, internal linking structures, and entity frameworks were automatically suggested to improve machine readability and enable better integration into the knowledge graph.
A key element was the governance approach: the platform prepared structures, interlinking, and formats, but did not publish anything without review. All content passed through the specialist departments and compliance approval. This allowed the institution to leverage the benefits of automation without violating regulatory requirements. In Phase 3, the new content was re-evaluated against ChatGPT responses and refined where necessary — for example, through more precise definitions, clearer trust signals, and more tightly interconnected specialist pages.
Results
After 90 days, the analysis showed a clear impact. The Semantic Authority Score rose from 36 to 64 points. In ChatGPT, the brand was mentioned in 19 out of 30 standard prompts, compared to 6 out of 30 previously. It appeared as the primary recommendation 11 times, up from 2. Across all five LLMs, share of voice increased from 12% to 31%.
The operational effects were equally measurable. Manual analysis time for the marketing team dropped from approximately 20 hours to 6 hours per week. The production time for new, approved topic clusters decreased from an average of 21 days to 9 days. Based on an internal hourly rate of €95, the calculated return on investment was reached after approximately 5 months. Particularly relevant for the institution: the improvements were achieved without any compromise to compliance, tone, or subject-matter accuracy.
Lessons Learned
Summary
With Zeno Visibility, the financial institution did not merely implement LLM brand monitoring — it directly translated the analysis into a semantically structured authority system. As a result, brand visibility in ChatGPT and other LLMs increased measurably. For regulated organizations, this case demonstrates that visibility in generative AI responses is not purely a monitoring challenge, but a question of content infrastructure.
Further Case Studies
---
*This content was created with AI assistance and reviewed by a human editor.*