Generative artificial intelligence has captivated the world’s attention with its potential to reshape industries and businesses.
For many financial services executives — including those in payments — the question is not whether generative AI will profoundly impact their industry, but rather how to harness its immense power to create value.
This cutting-edge form of artificial intelligence revolutionizes data creation, unraveling existing patterns to generate entirely new data samples. Imagine witnessing a generative AI system ingest images of animals, only to astonish with awe-inspiring, never-before-seen creatures born from its creative prowess.
While payment companies have long utilized AI to enhance their products and combat fraud, the game-changer lies in the emergence of the large-language models, or LLMs. These models empower machines to contextualize, infer and independently create diverse content.
Payments titan Stripe is using ChatGPT — which is the most well-known application of generative AI — to distill insights from developer documents. This empowers developers to spend more time building products instead of hunting for information. The fintech Klarna has partnered with OpenAI, the company that developed ChatGPT, to create highly personalized and intuitive shopping experiences using its GPT-4 plugin.
Users can receive curated product recommendations and shopping advice, along with links to purchase products through Klarna’s search and compare tool. These are examples of a few early adopters that are experimenting with generative AI to improve productivity and business operations.
Three Areas Where Applying Generative AI Pays Off
But every payment company has an opportunity to unlock the power of generative AI in three areas of the business:
1. Crush fraud: Harnessing the unparalleled processing capabilities of generative AI, companies can combat fraud like never before. By detecting elusive patterns and anomalies in vast data sets, generative AI trains itself to identify fraudulent events before they occur. Open-source generative AI models, such as generative adversarial networks, or GANs, empower fraud teams to generate synthetic data that closely mirrors real transactional data.
This iterative process enables the development of robust data sets that make it possible to identify “actual” versus “fraudulent” transactions. And GANs are just one example among many options. Fraud teams can also choose variational autoencoders, or VAEs, and autoregressive models. Each offers different trade-offs in training stability, sample quality, computational requirement, and interpretability.
Want to go deep on AI best practices for banks?
Attend our AI Masterclass — Unlocking the Power of Artificial Intelligence in Banking — at The Financial Brand Forum 2024 on May 20-22 in Las Vegas. Led by Ron Shevlin, chief research officer at Cornerstone, this three-hour workshop will be jam-packed with lessons learned from industry leaders and real-world case studies.
For more information and to register, check out the Forum website.
2. Empower CFOs: Generative AI’s transformative potential extends beyond streamlining operational processes, to empowering chief financial officers. By processing, summarizing, and extracting vital information from vast financial documents, such as annual reports and financial statements, generative AI equips CFOs with invaluable insights for more efficient analysis and decision-making. Expertly fine-tuned models generate diverse scenarios, simulating market conditions and economic indicators, enabling CFOs to predict future trends, asset prices, and potential risks and opportunities. With generative AI, finance operations can reach new heights of efficiency and strategic prowess.
3. Energize marketers: Generative AI empowers marketers to unleash the full potential of personalized marketing, engaging customers like never before. After studying customer behavior patterns, generative AI can create tailored content that resonates deeply with specific audience segments. For instance, if a customer interacts with a competitor’s product, generative AI can dynamically generate compelling content to promote the company’s own comparable offering to that customer.
Furthermore, generative AI equips sales teams with precise analytics and customer insights, enabling targeted marketing campaigns. Open-source platforms like ChatGPT, GPT-4, Bard, Alpha Code, Claude, and DALL-E 2 provide marketers with the tools to revolutionize content creation, automate processes, and accelerate growth.
But Beware of These Generative AI Risks
However, understanding the risks associated with this cutting-edge technology is crucial. Consider the following key risks when exploring generative AI use cases:
• Biased training data: The quality of the source data shapes the outputs of any AI solution. Biases present in human-inputted data can be amplified in generative AI outputs, leading to skewed results. Moreover, using sensitive or confidential data raises ethical concerns, and any errors in the source material can result in generating content that spreads misinformation.
• Loss of intellectual property: Generative AI requires vast amounts of training data, including confidential information and personally identifiable information. This data is retained by the model, potentially exposing private or proprietary information to other users outside the company. As more businesses adopt generative AI, the risk of unauthorized access to this data increases.
• Muted marketing content: Machine-generated content still lacks the emotional intelligence that comes with human creativity. As such, it can mute the personality of a brand, which, in turn, can negatively impact customer loyalty, engagement and brand perception by creating poor quality, inaccurate, or offensive content.
• Hallucination: Hallucination refers to the tendency of generative AI models to confidently produce inaccurate results, especially when trained on public data. Data sourced from the internet may include biased, outdated, incomplete, or inaccurate information, leading to the spread of misinformation. Ethical concerns arise when sensitive and confidential data is used, as it may also contribute to the generation of fake content.
Read more about generative AI:
- 4 Steps to Dodge Trouble When Using Generative AI
- The Best Way to Implement GenAI (and Avoid Random Acts of Digital)
- Notably Quotable: What Banking & Tech Leaders Think of Generative AI
Five Steps for Building a Generative AI Strategy
To get started with generative AI, payment companies should follow these actionable steps:
1) Develop generative AI literacy: Educate teams throughout the organization on the technology, use cases, and potential implications. Bringing in experts for training will foster a positive culture of responsible adoption.
2) Build a North Star for generative AI: Formulate a practical strategy and roadmap, aiming for a generative AI “North Star.” Identify areas within the business that can benefit from generative AI and determine desirable use cases while keeping the associated risks in mind. Designate a leader responsible for cross-enterprise coordination, driving efficiency, learning from pilots, prioritizing investments, and championing success.
3) Dive in, build and test: Embrace generative AI platforms that offer consumable models and applications through application programming interfaces, or APIs. Customize these platforms for specific use cases to achieve quick returns. Pretraining models by fine-tuning them with proprietary data will lay a solid foundation for progress. Early adoption will uncover resource and capability gaps, enabling the company to invest strategically in people, technology, and process changes.
4) Get technology platforms and data ready: Address the data challenge by adopting a strategic approach to acquiring, refining, safeguarding, and deploying data. Establish a modern enterprise data platform built on the cloud, providing a trusted and reusable set of data products.
5) Invest in people and technology: Allocate resources to develop technical competencies such as AI engineering, data modeling and architecture. Provide training across the organization to effectively work with AI-infused processes. Ensure the company has a modern payments platform that seamlessly connects to third-party platforms via APIs. Take a strategic and disciplined approach to acquiring, refining, and deploying data.
Generative AI is revolutionizing the payments industry, presenting endless possibilities. From streamlining operations, to intelligent fraud management, to converting information into insights, its impact will be felt across the entire business. By exploring and experimenting with generative AI today, payment companies can secure a future of accelerated growth and transformation.
About the author:
Safwan Zaheer is managing director at the consulting firm HundredX.