LLM Industry Benchmark 2026: Which Sources Shape AI Finance & Insurance Recommendations

AI assistants such as ChatGPT, Gemini and Google AI Overviews are increasingly answering commercial questions about bank accounts, brokers, loans and insurance before anyone visits a provider’s website. Which sources sit behind those answers is therefore a relevant question for banks, brokers and insurers. Peak Ace analysed 1,000 standardised prompts across ten industries to investigate this. This article presents the Finance & Insurance findings and places them in the context of the wider benchmark.

Note: This article is an excerpt from the LLM Industry Benchmark. The complete analyses, including results for all ten industries, are available in the benchmark itself.

In short:

• In this benchmark, comparison formats are retrieved most often for three of the four search intents in Finance & Insurance.
• For criteria and guidance questions, articles lead in finance, while how-to guides lead across all ten industries.
• The sources mix differs by industry, so the same type of question can draw on different source ecosystems.
• All findings refer to this prompt set, this time period and this setup.

Methodology and data corpus

The benchmark is based on a clearly defined data corpus. The following sections describe its structure and the key performance indicators (KPIs) derived from it.

Prompt set and structure

The prompt set comprises 1,000 standardised bottom-of-funnel prompts. There are 100 commercially relevant prompts per industry, distributed across ten topic clusters. The prompts are geared towards purchasing decisions, i.e., questions people ask shortly before opening an account, choosing a broker or taking out a policy. All prompts are brand-neutral, so no brand names appear in the questions.

Search intent

Each industry covers four search intents, with 25 prompts per intent:

• Best/Top/Test Winner: Recommendation questions, such as “Which broker is best for an ETF savings plan?”
• Comparisons: Contrasts between two offers, such as a fixed-term deposit versus a savings plan.
• Prices/Where to buy: Questions about costs and providers, such as what a current account abroad costs.
• Criteria/Guidance: Questions about selection criteria, such as “What should I look out for when choosing a broker?”

Market, language and time period

The prompts are in German and geared towards the German market. Answers were collected from 1 to 31 August 2026.

Systems and repetitions

ChatGPT, Gemini and Google AI Overviews were observed. Each prompt was executed 15 times, five times per system. Individual responses were considered samples, which allows for a better classification of fluctuations.

Database

The data comes from Peec AI exports. Fifty brands per industry were monitored, and the report focuses on the top 20 brands by visibility.

Key figures

• Visibility: Shows how frequently an entity appears in the responses to the prompt set.
• Share of Voice (SOV): Describes the proportion of an entity within the top 20 of the industry.
• Sentiment: Rates the tone of the responses that mention an entity on a scale from 0 to 100, where higher values are more positive.
• Retrieval: Means that a source was accessed as an information base. It isn’t the same as an explicit citation in the final answer.

Search intents and source patterns

Search intent appears to influence the type of source that shapes a response. The same type of question encounters different mixes of sources across industries. In this benchmark, the Finance & Insurance mix differs from the cross-industry picture in several places.

Finance & Insurance compared with all ten industries

The cross-industry column is aggregated and doesn’t refer to a single industry.

Four patterns stand out in the Finance & Insurance data:

• Comparisons lead in three out of four intents. They appear not only in explicit “A versus B” questions, but also in recommendation and pricing questions.
• Guidance changes the mix of sources. For criteria and guidance questions, articles lead in finance, followed by comparison pages and how-to guides. Across all ten industries, how-to guides account for 22.33 per cent of retrieval in this intent, which points to a different starting point.
• Product pages sit alongside comparisons. Product pages rank among the top three source types for recommendation, comparison and pricing questions. A product-page classification alone doesn’t show whether a page belongs to a provider or to a third party.
• Listicles don’t feature among the leading formats. No Finance & Insurance intent places listicles in its top four source types. This doesn’t mean zero retrievals; it means they aren’t a leading format in this dataset.

The pattern isn’t universal. Finance follows the cross-industry lead for recommendations, comparisons and pricing, but not for criteria and guidance. This differs by industry, which is why the industry serves primarily as a lens for reading the data, not as a scoreboard.

Retrieval means that a source has been accessed. This doesn’t automatically imply that it will be cited or that the brand will appear in the final answer. The URL types describe formats, not ownership or authority.

AI Visibility Focus

An average value doesn’t show whether mentions are spread across many brands or concentrated on a few. The analysis therefore considers two metrics within the top 20 of an industry. Both are based on SOV shares normalised to 100 per cent:

• HHI: Describes the concentration of SOV components. Higher values indicate a more uneven distribution.
• Top 3 share: The combined share of the three largest SOV contributions.

In Finance & Insurance, the three largest SOV contributions account for 40 per cent of the normalised top 20 share, with an HHI of 856. This indicates a moderate concentration signal for this setup. It describes how mentions are distributed within the observed top group, not market shares, and it says nothing about demand or revenue distribution.

In the Finance & Insurance prompt set, ING (19 per cent visibility), Trade Republic (15 per cent) and Scalable Capital (12 per cent) show the highest visibility in the observation period. All three are banking or brokerage brands; no insurer appears in the top three. This may indicate that for purchase-near questions about accounts and brokers, banks and brokers are often directly relevant to the answer. This is a snapshot of this setup, not a statement about market share. Visibility isn’t an endorsement of a provider or evidence of the suitability of its products.

Concentration scores help to understand the distribution of mentions. However, they don’t explain why a particular provider appears. For that, it’s worth looking at the questions and the pages accessed behind your own results.

What does this mean for SEO, Digital PR and marketers in finance?

The data reveals a pattern: depending on the intent, answers in this benchmark draw on different types of sources. This can help banks, brokers and insurers position themselves more deliberately. The goal isn’t to be visible for every question, which is hardly realistic. It’s to understand what you offer, where you have well-founded answers, and where visibility is worth the effort. The results point to the following considerations:

• Handle answer types separately. A comparison should explain alternatives, including their advantages and disadvantages. A guide should clarify requirements or support a decision. A general article rarely fulfils both tasks equally well. Practical implication: map each product line to the intent and format it should answer.
• Align prices and terms across sources. Comparisons are central to pricing and where-to-buy questions. Check whether fees, interest rates, contract durations, cancellation policies and eligibility criteria are described identically on your product pages, overview pages and in relevant comparisons.
• Explain the criteria and link them to offers. Articles and guides lead for criteria questions in finance. Explain selection criteria clearly, such as costs, conditions and the differences between offers, and link them to suitable products so that advice and offer don’t drift apart.
• Prioritise your Digital PR. Compile a list of the most important third-party sources per intent and check what they say about your brand. Not every publication is equally relevant. The white paper names Finanztip, Finanzfluss, test.de and Verivox as examples of comparison and advice sources in finance. These illustrate the information ecosystem; they aren’t a measured ranking of leading domains.
• Keep information consistent. Product names, fees, conditions and terms should match on your website and in relevant external sources. Outdated or conflicting information increases the risk of inaccurate statements in generated answers.
• Don’t forget the SEO basics. Key content should be accessible, clearly structured and easy to navigate. This is a prerequisite for information access, but not a guarantee of recommendations in AI-generated answers.
• Treat monitoring as an ongoing task. Brand mentions and retrieved sources can be observed separately, ideally with a consistent prompt set. Analysis by category, intent and system complements traditional search and conversion metrics.

Limits of expressive power

The results are valid only for this specific prompt set, time period and setup. They don’t represent a complete measurement of an industry’s or brand’s visibility in AI systems. Specifically, the following should be noted:

• The prompt set shapes the outcome. Different questions or wording can lead to different answers.
• Individual responses are samples. Small differences shouldn’t be overinterpreted.
• Three systems aren’t one system. ChatGPT, Gemini and Google AI Overviews can’t simply be equated, so patterns may differ by engine.
• Retrieval isn’t citation. A fetched source isn’t automatically cited or mentioned in the final answer, so the sources mix can’t be read as a causal explanation of recommendations.
• Dialogue and personalisation are missing. Typical purchasing situations, including follow-up questions and personal circumstances, can’t be reliably described on this basis.
• Sentiment isn’t a quality assessment. It describes the tone of the responses in the dataset, not customer satisfaction with a provider.

Read the full benchmark

This article provides an excerpt. The complete LLM Industry Benchmark includes, among other things:

• The source mix for each of the ten industries and all four intent types
• Visibility, SOV and sentiment within the benchmark
• HHI and top 3 share within the top 20 of each industry
• The leading entities observed per industry
• The cross-industry shares of the source types
• The industry scorecards and the complete methodology

Download the full LLM industry benchmark.

Had a look at the benchmark and want to strengthen your telecoms brand’s presence in AI answers? Get in touch with Peak Ace’s LLMO and GEO team. 

FAQ: AI Visibility, Benchmarks and GEO in Finance & Insurance

What is a benchmark?

A benchmark evaluates defined results against a fixed test set. The LLM Industry Benchmark examines brand mentions, responses and retrieved source types for commercial questions across ten industries.

What does the LLM industry benchmark measure?

It captures visibility, SOV, sentiment, concentration and retrieval patterns for a defined prompt set, market and time period. The values are aggregated across ChatGPT, Gemini and Google AI Overviews.

What is AI visibility?

AI visibility describes how frequently a monitored brand or entity appears in responses to a defined set of prompts. This is distinct from whether a source is cited or quoted within the response.

Which sources do AI systems retrieve for finance questions?

In this benchmark, comparison formats are retrieved particularly often for recommendation, comparison and pricing questions, while articles lead for criteria and guidance questions in Finance & Insurance. This applies only to the examined prompt set and time period.

What is a GEO strategy?

Generative Engine Optimisation (GEO) combines SEO, content, Digital PR and monitoring to improve visibility in AI-generated answers. Measures are best prioritised according to the questions your target audience asks and the sources they are likely to encounter.

Does AI visibility depend solely on one’s own website?

Comparisons, articles, guides and other third-party formats play a significant role in this benchmark. It’s therefore worth checking specific URLs to see how your own content compares with relevant third-party sources.

How can you test whether measures are effective?

In three steps: a source and content audit, a check of the timeliness and consistency of the information, and a remeasurement with the same test questions under the same conditions.