The AI Boom Meets the Acquiring Wall: Why AI Startups Struggle with Merchant Account Approvals

CatalystPay Point of View, Stanimira Radeva

AI is everywhere. From chatbots to generative design tools, autonomous agents to deep learning SaaS platforms, artificial intelligence has officially gone mainstream. In 2025, it’s hard to find a digital business that isn’t embedding AI into its core offering. The phrase “AI-first” is more than just a trend; it’s becoming the default across industries.

And yes, full transparency - this very article is proudly AI-assisted (thanks, ChatGPT!). Because when we talk about the rise of AI, we’re living it too. 

Yet amid all the momentum, one unexpected roadblock continues to surprise AI founders: getting approved for a merchant account.

Let’s be clear: not all AI startups are considered high-risk, and not all acquiring banks are shutting their doors. Many AI-based businesses, from productivity & design tools to infrastructure providers, are already thriving with stable payment processing in place. 

But here’s the catch: acquirers don’t invest, they assess. And in the current environment, where investor enthusiasm has led to a wave of underprepared applicants, acquiring banks are more cautious than ever. They’re not judging vision, they’re judging risk.

At CatalystPay, we work with acquiring banks across the UK and EU to help merchants access payment processing. And increasingly, we see AI startups get flagged, not because they aren’t legitimate, but because their business model, security practices, or risk posture raise too many unanswered questions.

Here’s why that happens  and how AI-driven businesses can better prepare.

AI May Be the Future, But It’s Also Often a High-Risk MCC Today

To an acquiring bank, risk is not about hype or potential. It’s about stability, transparency, and liability. Many AI startups, particularly those offering SaaS, AI agents, image/video generation, or automated decision-making tools typically might be classified as high risk MCC as they raise red flags across all three.

Here are the key concerns that acquiring banks typically flag:

1. Unclear or Unproven Business Models

Many AI businesses launch with a “build fast, monetize later” mentality. From a product standpoint, that’s fine. From a payments perspective? It’s a red flag.

Banks and payment providers are under pressure to understand what is being sold, to whom, and how. When onboarding an AI business, they need to assess:

• What exactly are customers paying for?
  
• Is the output (e.g. AI-generated image or text) considered digital content, intellectual property, or a service?
  
• Are users charged per interaction, per subscription, or by API usage?
  

When business models are vague or changing (think: from API access to marketplace to SaaS to enterprise), it becomes harder for acquirers to categorize the risk. And when they can’t classify it, they often say no.

2. Chargeback Exposure and User Disputes

AI tools are increasingly used for automation, productivity, content creation, and even financial decision-making. But this opens the door to customer dissatisfaction, buyer remorse, and higher chargeback rates - especially in consumer-facing tools.

Examples we’ve seen:

• A user pays for an AI image generation tool but is unhappy with the result. They issue a chargeback claiming the product was “not as described.”
  
• An AI agent completes a task incorrectly (e.g. books a wrong flight or drafts the wrong legal doc). The customer disputes the charge.
  
• Subscription fatigue sets in after a user tries a flashy AI tool, forgets to cancel, and disputes the recurring charge.
  

These scenarios create operational risk for acquirers, especially when the AI’s “output” is subjective or unpredictable.

3. Deepfake & Content Moderation Risks

AI businesses working with image, video, or voice generation tools fall into a particularly sensitive category. Why? Because their platforms can be used (intentionally or not) to generate:

• Deepfake media
  
• Synthetic celebrity content
  
• Political impersonations
  
• NSFW or age-restricted material
  

Even if the startup includes filters or policies, the acquirer assumes potential liability if illegal or harmful content is produced and monetized.

Acquiring banks are highly risk-averse when it comes to digital content businesses that don’t fully control the end-use. And with AI, it’s often difficult to trace or guarantee content safety at scale.

4. Systemic Vulnerabilities and Security Concerns

The recent open letter from JPMorgan’s Chief Information Security Officer says it bluntly: “The way we use SaaS and AI is fundamentally broken.” Acquirers are listening.

Key concerns include:

• Direct API integrations into sensitive systems, often without proper access controls
  
• AI agents acting with elevated privileges across customer environments
  
• Lack of auditability -  i.e. can the startup prove what the AI did, and why?
  
• Data flow opacity  - where is customer data going, and is it being stored, fine-tuned, or shared?
  

When security is an afterthought, or when the company relies entirely on third-party AI providers (e.g. OpenAI, Anthropic, etc.), acquirers worry about systemic concentration risk. If the AI layer fails, gets breached, or is abused — who is liable?

5. Regulatory Grey Zones

In the UK and EU, data privacy and AI-specific regulations are advancing fast. GDPR already limits how companies can use personal data in automated systems. The upcoming EU AI Act will impose stricter obligations on companies building or deploying “high-risk AI systems” such as those used for biometric analysis, credit scoring, or behavioral profiling.

Acquirers don’t want to be caught in the middle if an AI startup is:

• Using training data without proper rights or consent
  
• Failing to offer transparency around how decisions are made
  
• Selling tools that could be classified as high-risk under regulation
  

Until there's more legal clarity, AI merchants are often treated as “better safe than sorry” rejections.

6. Subscription Business Models and the VAMP Factor

One area under sharper scrutiny lately is subscription-based AI services, especially in the context of Visa’s updated VAMP (Visa Acquirer Monitoring Program) rules. Under VAMP, acquiring banks face tighter obligations to monitor merchants with elevated risk profiles, particularly those with recurring billing models, high chargeback ratios, or opaque customer experiences.

For AI startups offering monthly subscriptions to image generation platforms, chatbot tools, AI agents, or productivity SaaS, the billing mechanics often fall right into VAMP-sensitive territory.  

Lack of clear cancellation paths or trial-to-paid conversion disclosures, inconsistent refund practices, customers surprised by recurring charges from a service they barely used, AI-generated outputs that can’t be disputed or returned, leading to “not as described” claims - just few of the potential flags in the eyes of the acquirer.

Acquirers know the playbook: subscription fatigue + unclear deliverables = chargeback spike. And with VAMP in place, banks now face pressure to either justify the risk or offboard merchants preemptively to avoid penalties.

Even legitimate startups with honest intentions can get caught in this pattern. If they lack proper descriptors, billing transparency, or dispute-handling policies, they may trigger acquirer alerts before they even scale.

The Acquirers' Lens: What They Look For (and Often Don’t See)

At CatalystPay, we work closely with both merchants and acquiring partners. Based on the merchant onboarding of tens of AI business cases, here’s what acquiring banks look for in AI companies  and where they often find gaps:

What Acquirers Want	What AI Startups Often Lack
Clear product/service description	Vague or evolving use cases
Transparent pricing model	Complex or dynamic billing logic
Defined target audience	Open platform with unknown users
Evidence of strong T&Cs / refund policy	Thin or missing documentation
GDPR / data compliance clarity	Unknown training data or unclear data flow
Responsible use policies	Loose control over how AI is used by end users
Proof of chargeback handling	No dispute handling process or past data

 

This doesn’t mean AI startups are unqualified. It means they need to prove they’re payment-ready, not just product-ready.

What an AI Startups Should Consider when Applying for Merchant Accounts

Want to increase your chances of approval? Here are some steps we recommend:

1. Define Your Commercial Model in Plain English: What are customers buying, exactly? Spell it out. Describe your product like you would to a bank examiner.
   
2. Have Strong Terms of Service and Refund Policies: Show acquirers you’ve thought about consumer protection. Include refund windows, content policies, and responsible use clauses.
   
3. Create a Clean KYC Package: Include all documents upfront: company registration, director IDs, domain ownership, business model explanation, and product screenshots.
   
4. Clarify Data Handling: Are you using OpenAI’s API? Are you storing user inputs? Training models on customer data? Say so clearly and explain how you stay GDPR-compliant.
   
5. Acknowledge the Risk and How You Mitigate It: If you know your tool could be used for questionable content, show your filters or manual review process. If AI decisions affect money or outcomes, explain your fail-safes.

How CatalystPay Can Help

We’re not just a middleman, we’re your payments ally. At CatalystPay, we work with a network of acquirers across Europe and the UK to find the right match for your business. We know how to position AI startups responsibly and transparently, even when your business model doesn’t fit neatly into a traditional MCC box.

We help you:

• Understand what documents and controls acquirers need
• Navigate industry-specific risk perception
• Strengthen your onboarding package and KYC story
• Connect with acquiring partners open to innovative business models

Final Thoughts: Not All AI Startups Are High Risk, But the Bar Has Risen

Let’s circle back to where we started: AI isn’t inherently high-risk, but the way it’s being deployed (fast, loose, and under pressure to impress investors) often raises concerns for acquiring banks.

The issue isn’t incompatibility. It’s unpreparedness. Too many AI startups jump on the hype train before laying the operational groundwork that acquirers require: clear models, strong policies, and transparency across the board.

At the same time, many AI-based businesses are already processing payments reliably, proving that it can be done, and done well.

At CatalystPay, We believe payments should keep up with innovation and that includes supporting the next generation of AI-driven businesses. But it also means AI businesses must meet a higher standard, not just pitch-deck ready, but payment-ready.

If you're building the next big thing in AI and need a partner who understands both innovation and risk — we're here to help. At CatalystPay, we work with AI startups to navigate onboarding, find the right acquiring partners, and set up payment infrastructure that scales with confidence.

Ready to get started? Apply for a merchant account here