AI in payments and fintech: Enhancing human decision-making and innovation

It’s 2025, and AI continues to be one of the most transformative (and talked-about!) forces shaping the payments and fintech landscape. From credit modeling and fraud detection to real-time customer service and underwriting, AI is already fundamentally changing how financial services operate. However, one key theme emerging from two recent industry event discussions is that AI is not replacing humans at all. It is enhancing their ability to help make better decisions and improve operational efficiency.

At Marqeta’s recent Embedded Edge event and the FTT Fintech Lending event in London, industry leaders came together to discuss the evolving role of AI in payments and lending. Experts including those from Lendable, Allica Bank, GDS Link, and Jaja Finance, shared their perspectives on how AI is already driving tangible impact, from enhancing credit decisioning and fraud prevention to improving customer experience and operational efficiency. This blog captures their key insights and commentary on how AI is reshaping financial services today.

One of the most significant use cases of AI in fintech is in credit modeling. Traditional credit scoring models rely on static data that is weeks out of date, often leading to outdated risk assessments. Today, AI and machine learning are continuously improving these models by dynamically updating credit scores based on real-time data. Open banking is helping here too.

Lendable, for example, is leveraging AI to enhance credit decisioning while maintaining human oversight for complex cases. 

Machine learning can update risk models as required - Jaja Finance mentioned they do so every 6 months, refining decisions based on evolving borrower behavior.

Experts agreed that while AI can and should handle straight-through lending decisions, human expertise remains essential for complex cases where context matters.

But what happens when AI becomes so sophisticated that it starts dictating which lender a customer should choose? Lendable raised this thought-provoking question, noting that if AI-driven lending decisions become fully commoditized, brand differentiation could diminish. To stay ahead, lenders must consider how they can maintain a unique value proposition even as AI continues to streamline decision-making.

AI is revolutionizing customer service in fintech, making it faster and more efficient than ever before. Companies are deploying large language models (LLMs) to power AI chatbots that can resolve customer queries in seconds rather than minutes.

At the FTT Lending event, one example stood out: AI-powered chatbots can now answer customer inquiries in an average of nine seconds, compared to the three minutes it typically takes a human agent. Jaja Finance told us that AI-driven solutions handle roughly 50% of customer service queries, allowing human agents to focus on more complex issues that require empathy, discussion and critical thinking.

Yonder, a fintech prioritizing quality over quantity in recommendations, echoed a similar sentiment. While AI can augment decision-making, humans still play a crucial role in curating experiences. Just as Apple News and Spotify rely on human editors to refine recommendations, Yonder blends AI-driven insights with human judgment to ensure high-quality guidance and recommendations. 

Fraud remains a major challenge in the financial sector, but AI is proving to be a powerful ally in combating financial crime. AI-driven fraud detection tools are now capable of screening and validating documents with unprecedented accuracy.

One example from the FTT Lending event highlighted an AI fraud tool that detects document manipulation, such as altered PDFs, helping businesses prevent fraudulent loan applications. For Allica bank, this tool alone is currently identifying “over £1 million per week in fraudulent applications”, a testament to the power of AI in safeguarding financial institutions from bad actors.

In the business lending space, AI is also being used to analyze financial documents with greater efficiency. By scanning and summarizing lengthy contracts or surveyor reports, AI can extract key insights and surface relevant information, significantly reducing the time and effort required for due diligence that a human can then use to make faster lending decisions. 

The panelists agreed that one of the most exciting applications of AI in fintech is in underwriting. Liberis told us that they consider themselves at the forefront of this transformation, exploring how generative AI (GenAI) can optimize sales calls, improve customer interactions, and enhance risk assessment.

For example, AI can analyze bank statements using LLMs, providing more accurate financial assessments and improving underwriting algorithms. Additionally, Liberis is experimenting with AI-powered underwriting agents that leverage GenAI models like Gemini to assist human underwriters in making more informed decisions.

AI’s influence in payments extends beyond credit modeling and fraud detection—it is also enhancing spend management and corporate payments. One of the key discussions at our Embedded Edge event focused on how AI can bring more granular visibility into corporate card spending.

For example, AI-driven spend management solutions can track inter-departmental expenses close to real-time, flag anomalies, and automate expense approvals. This not only reduces the administrative burden but also helps businesses optimize their financial operations.

AI-powered smart payment cards are emerging and promise to change the game even further. These cards offer customizable spending controls, real-time transaction monitoring, and AI-driven recommendations to ensure optimal payment decisions. As businesses navigate the complexities of multi-currency and multi-jurisdiction payments, AI can help streamline international transactions, ensuring compliance with local regulations while optimizing costs.

Despite AI’s advancements, one clear theme emerged from all these discussions: AI is here to help enhance human decision-making, not replace it.

For lending decisions, AI can provide underwriters with better data, but humans will continue to make complex judgment calls. For customer service, AI-powered chatbots can handle routine inquiries, but human agents will remain essential for nuanced support. For fraud prevention, AI can detect anomalies at scale, but fraud investigators will still play a vital role in interpreting complex cases.

Ultimately, AI is not a replacement for expertise. It’s a tool that empowers fintechs to operate more efficiently, make smarter decisions, and deliver better customer experiences for businesses and consumers alike.

As AI technology continues to evolve, its impact on payments and fintech will only grow. Companies that successfully integrate AI into their operations will gain a competitive edge by offering faster, more personalized, and more secure financial services.

However, to fully harness AI’s potential, fintechs must strike the right balance between automation and human oversight. They must also stay ahead of evolving regulatory frameworks to ensure responsible AI deployment.

At Marqeta, we believe that AI’s role in fintech is about augmentation, not automation. By combining AI’s speed and efficiency with human expertise and judgment, the financial industry can build a future that is both innovative and customer-centric. As we move forward in 2025 and beyond, the fintechs that embrace AI as a collaborative tool will be the ones that shape the future of financial services.