Inside the tool
A real BuzzView analysis in this category
Solaris’s AI visibility across ChatGPT, Google AI Overviews & Perplexity — one of the brands tracked in this category, straight from the live tool.
We sent 48 real buying questions to ChatGPT and Google AI Overviews – the questions fintechs and founders actually ask. Then we counted which providers get named, and how often. The result is a ranking built from real AI answers, not opinion.
Stripe (3.2%) and Qonto (2.4%) dominate the answers – the remaining 235 named providers split what's left. If you're not here, you simply don't exist to AI users. That's exactly the gap BuzzView makes visible.
Share of all brand mentions across 48 prompts (Share of Voice). The longer the bar, the more often AI names the provider – across every question tested.
48 prompts. 235 providers. 496 brand mentions. Here is what the distribution actually means — for the category, and for every provider in it.
Share of Voice across 496 total brand mentions — top 6 vs. remaining 229 providers.
The banking and embedded finance category is one of the most fragmented B2B software niches in the DACH market. When AI receives a buying prompt about banking-as-a-service or embedded payments, it draws from a pool of 235 distinct providers — a remarkable number that reflects a decade of fintech infrastructure build-out. Every sub-segment of this market — core banking modernisation, card-as-a-service, embedded lending, B2B payment facilitation, revenue-based financing — has attracted its own cohort of specialist providers, and AI has learned enough about all of them to name at least some of them some of the time.
Yet the distribution of those 496 mentions follows a severe power law. The top 6 providers — Stripe, Qonto, Billie, Solaris, Mambu, and Banxware — collect just 14% of all mentions. That sounds small until you consider what it means in reverse: the other 229 providers collectively share 86% of a very thin remainder. The average non-top-6 provider appears fewer than 2 times across all 48 prompts — a level of presence so low that no realistic buyer decision is being influenced by it. The fragmentation that defines this market at the product level carries over exactly into AI recommendation behaviour.
This creates a paradox for embedded finance providers. The category is broad enough that AI systems encounter dozens of contextual cues when answering any banking-related prompt, yet narrow enough that only a handful of brands have accumulated the depth of online signal — documentation, analyst coverage, comparison content, community discussion — that makes AI confident enough to name them repeatedly. A provider that appears in 5 sources with 200 words each is not the same, to an AI, as one that appears in 50 sources with 2,000 words each. Depth and breadth of third-party corroboration is what converts occasional name-dropping into consistent recommendation.
The implication for any B2B fintech in this space is direct. The 235-provider pool is not a competition for share of voice so much as a competition for existence. Getting named at all is the first gate. Getting named consistently, across different prompt framings, is the second. Getting named with positive framing is the third. Most of the 235 clear none of these bars — they appear once or twice across highly specific prompts and then vanish. Building a path from occasional mention to consistent recommendation requires treating AI visibility as a deliberate content programme, not a side-effect of other marketing activity.
In a 235-provider field, being named is not enough. Consistent, cross-prompt presence is what separates the 6 from the 229 — and that consistency is built, not inherited.
Cross-category visibility scores: Solaris 100%, Billie and Mambu 87.5%, Re:cap 25%.
Visibility, as measured in this analysis, is the share of the relevant prompt sub-categories in which a provider appears at least once. Solaris scores 100%: it gets named regardless of whether the prompt is about Banking-as-a-Service for SaaS companies, core banking technology for financial institutions, B2B payments, or embedded lending for non-bank lenders. That cross-category presence is not accidental. Solaris has spent years publishing detailed technical documentation, contributing to developer community discussions, securing analyst mentions across multiple fintech research reports, and building integration partnerships that generate secondary coverage across partner channels. The 100% visibility score is the outcome of that accumulated signal, not a reflection of any single campaign.
Building consistent visibility across prompt types requires a fundamentally different content strategy than ranking for a single keyword cluster. AI systems synthesise across many sources when building their answer — product documentation, third-party reviews, press coverage, technical blog posts, industry analyst reports, and community forum threads. A provider that is well-documented in one area but sparse in others will appear when the prompt matches its strong suit and disappear when it does not. The AI is not choosing to ignore a provider in a given sub-category; it simply has insufficient corroborating material to include it with confidence. Silence in the training data produces silence in the output.
The visibility spread among tracked providers illustrates this precisely. Billie and Mambu reach 87.5% — present in 7 of 8 sub-categories — while Re:cap scores only 25%, appearing in just 2 of 8 prompt types. The gap is not primarily a product quality gap. Solaris and Re:cap may both be credible infrastructure providers serving defensible market positions. The difference is the breadth and depth of publicly available documentation that AI can draw on when constructing an answer. Re:cap's revenue-based financing product is real; its online footprint in adjacent sub-categories — such as core banking, B2B payments, or Banking-as-a-Service — is thin enough that AI cannot confidently include it in answers to prompts framed around those topics.
For providers currently below 75% visibility, the diagnostic question is not "why doesn't AI recommend us" but "which prompt sub-categories are we invisible in, and what is missing online that would let AI include us in answers to those questions?" The answer is almost always content: case studies in missing verticals, technical documentation for underrepresented use cases, or third-party editorial mentions in categories where the provider has a product but no public signal. Visibility gaps are content gaps — they are diagnosable and they are fixable, but only if the gap is measured in the first place.
Visibility at 100% is not luck. It is the result of having documentable, third-party-corroborated presence across every sub-category a buyer might search in. The gaps in a visibility score show exactly where to build.
Four prompt intent types were tested — each surfaces a different subset of the 235 providers.
The 48 prompts in this analysis cover four distinct intent categories: best-of discovery ("what are the best Banking-as-a-Service tools for SaaS providers in Germany"), head-to-head comparison ("compare Solaris, Kontist and Stripe in the area of embedded financial services"), alternative-seeking ("what are alternatives to Kontist for embedding financial services into products"), and use-case filtering ("best B2B payment tools for e-commerce merchants in Germany"). Each of these intents pulls from a different slice of the AI's training signal, and the providers that win in each are not identical. A provider can dominate discovery prompts while being invisible in comparison prompts, or vice versa — and most providers in this category have a strong imbalance across the four types.
Discovery prompts — "what are the best X" — tend to surface the providers with the broadest general presence: those named in "top 10" lists, analyst round-ups, and high-traffic comparison platforms. Stripe and Qonto dominate here because they appear in the greatest number of general overview resources. A company that has never appeared in a category round-up or an industry overview article is extremely unlikely to appear in a discovery prompt response, regardless of product quality or market fit. Discovery visibility is essentially a function of how many independent third parties have decided to name you when writing about the category as a whole.
Comparison prompts — especially those naming specific providers — produce more nuanced responses. When a buyer asks to compare Solaris, Kontist, and Stripe, the AI must draw on structured, attribute-level information: pricing transparency, geographic availability, API documentation quality, integration ecosystem breadth, regulatory licensing. Providers that invest in transparent, detailed technical documentation tend to outperform in comparison contexts even when their general discoverability is lower. The AI can only say something substantive about a provider's strengths if the training data contains substantive, specific information about them.
Use-case and vertical filtering prompts are where newer or more specialised providers can break through against incumbents. Billie's strong showing — 2.4% SOV overall, but 87.5% cross-category visibility — is partly explained by this dynamic: its B2B BNPL positioning is specific enough that when the prompt context matches its primary use case, it surfaces consistently even against much larger, more broadly known competitors. The lesson for specialist providers is not to try to compete with Stripe in discovery prompts, but to dominate the specific use-case and vertical prompts that match their actual product positioning — and to build the content that makes that dominance possible.
Winning across prompt types requires three distinct content layers: broad overview articles for discovery, structured technical specs for comparison, and specific vertical case studies for use-case queries. Most embedded finance providers invest in one and neglect the other two.
Positive-mention rate: Mondu 67%, Billie 42%, Banxware 44%, Adyen 13%, DATEV 0%, Temenos 0%.
Not all brand mentions are equal. In this analysis, every AI-generated mention is classified as positive (the AI explicitly recommends or endorses the provider in context), neutral (named without qualitative framing — present in a list but not distinguished), or negative (named with cautions, limitations, or criticism). The sentiment distribution reveals a second dimension of AI visibility that raw mention count completely obscures. A provider can score 8 mentions and zero positive responses; another can score 9 mentions with 6 positives. From a buyer's perspective, these are not comparable outcomes — and from a commercial perspective, only one of them is doing useful work.
Mondu leads the tracked set in positive-mention rate: 6 of its 9 mentions carry an explicit positive framing. Billie follows with 5 positives from 12 mentions. Banxware earns 4 positives from 9 mentions. At the other end, DATEV and Temenos each accumulate 7 mentions with zero positive sentiment — every single mention is neutral, meaning AI names them but offers no endorsement whatsoever. Adyen, despite 8 total mentions and a global brand presence, generates only 1 positive response in this context. The divergence is stark: some providers are being named and recommended; others are being named and left to the buyer to evaluate without guidance.
This matters because the buyer intent behind most of these prompts is evaluative. The user is not asking "who exists in this space" but "who should I use, consider, or at least look at seriously." An AI response that lists DATEV neutrally alongside Billie with explicit positive framing creates a materially different conversion outcome for each brand. The buyer reading that response infers something about relative standing even when the AI has not made an explicit comparison. Positive framing in an AI answer functions as a soft endorsement — and soft endorsements from AI systems are increasingly the first filter in a B2B buying process that begins with a prompt rather than a search query.
The sources that drive positive sentiment are distinct from those that drive neutral mentions. Neutral mentions typically originate from list articles, Wikipedia-style overviews, and general category summaries — resources that aggregate names without evaluating them. Positive sentiment tends to originate from customer review platforms (G2, Capterra, Trustpilot), published case studies with measurable outcomes, editorial coverage that draws explicit conclusions about performance, and community discussions where practitioners recommend specific tools. A provider that has built a large list presence but thin review coverage will skew neutral; one that has fewer list appearances but strong, outcome-focused review coverage will generate fewer total mentions but a much higher proportion of positive ones. Positive sentiment is harder to build than raw visibility — it requires real customer success stories, documented outcomes, and third-party editorial credibility — but it is also harder for competitors to replicate quickly, making it the most durable differentiator available in an AI recommendation context.
Neutral mentions get you named. Positive mentions get you chosen. Building positive AI sentiment requires customer evidence, editorial credibility, and outcome-focused content — not just category presence.
235 providers named, 14% concentration in the top 6, median tracked visibility ~62% — a category where the hierarchy is still forming.
Taken together, the numbers that define this category — 235 providers named, 14% concentration in the top 6, average SOV for tracked providers under 2%, and a visibility median around 62% — point to a category that has not yet settled into a stable AI recommendation hierarchy. Compare this to more mature software categories such as project management, CRM, or email marketing tools, where 3 to 5 providers consistently capture 40 to 60% of AI mentions and the ranking has been largely stable across retraining cycles. Banking and embedded finance has not reached that equilibrium. The hierarchy is still forming, and the gaps between current positions and achievable positions are unusually large relative to the investment required to close them.
The reasons are structural. The embedded finance category is relatively young: most of the infrastructure providers that exist today were founded after 2015, and several of those tracked here — Banxware, finmid, Upvest — after 2018. That means the body of third-party content about them is substantially thinner than for mature SaaS categories. Analysts, journalists, and comparison platforms have not yet built the same depth of cross-provider coverage that exists for, say, HR software or accounting tools, where decades of competitive market development have generated thousands of comparison articles, review entries, case studies, and community discussions. The AI has less material to work with, which is why the category produces 235 provider names but concentrates meaningful mention share so narrowly among those with the deepest documentation.
There is also a product complexity factor. Embedded finance infrastructure is harder to explain clearly than a typical SaaS product. API-first banking-as-a-service, card-as-a-service, embedded lending rails, and B2B payment facilitation each require substantial technical context before a reader — or an AI — can meaningfully compare providers on relevant dimensions. That complexity raises the content investment required to achieve AI visibility. It means that providers willing to invest in detailed technical documentation, developer-oriented case studies, integration guides, and multi-format educational content will outpace those that rely on standard marketing copy alone. Standard marketing copy does not give AI systems enough structured information to form a confident opinion; technical depth does.
The openness of the current ranking also means that positions are not locked. A provider currently at 50% visibility and 1.4% SOV is not structurally disadvantaged in the way that a page-2 SERP ranking used to be — where domain authority accumulated over years and was nearly impossible to displace quickly. AI recommendation hierarchies in new categories are being formed now, from whatever content exists online, and they are updated continuously as models are retrained. The relevant benchmark for a European embedded finance provider is the cluster from Solaris to Banxware: 75 to 100% cross-category visibility and 1.8 to 2.4% SOV. That range is achievable without global scale. And it is the zone where AI recommendations start generating consistent inbound deal flow — the threshold above which being named translates reliably into being considered.
The AI recommendation hierarchy in Banking & Embedded Finance is still forming. Providers that build comprehensive, technically detailed, multi-format content now will establish positions that compound for years — before the hierarchy closes and displacement becomes genuinely hard.
No wishful thinking: the ranking comes from exactly these prompt types – best-of questions, comparisons, alternatives and use cases.
Visibility score = share of prompts where the provider appears in the AI answer at all. 100% means: present for every relevant question.
See it in action
Track your brand across every AI — automatically. Here's how it works.
1. What is BuzzView?
2. Your Brand vs. Competition
3. From Data to Action
Solaris’s AI visibility across ChatGPT, Google AI Overviews & Perplexity — one of the brands tracked in this category, straight from the live tool.
We set the real search and buying questions of your industry – exactly how your customers actually ask AI.
Every prompt runs against all major AI models. We count mentions, position, sentiment and the cited sources.
You see your ranking, your share of voice and exactly the prompts where competitors win – and you don't.
Run your own industry comparison and see in minutes whether ChatGPT & co. recommend you – or your competitors.
Start Free Trial No credit card · Results in minutes · GDPR-compliant, hosted in Germany
