3 Tips for Credit Unions to Get Started With GenAI Implementation
GenAI’s future role within FIs hinges on strategic integration with existing infrastructure and best practices.
The McKinsey Global Institute estimates that generative AI (GenAI) could add the equivalent of $2.6 trillion to $4.4 trillion in value to global industries each year. Banking leads those industries, to which McKinsey estimates GenAI has the potential to deliver between $200 and $340 billion annually.
Unlike other forms of AI, GenAI produces content based on prompts and directions from a person and the ideas or data produced are meant to resemble or enhance original human-generated work.
When applied to a tool such as a chatbot at a credit union, GenAI can analyze transaction data and member behavior to personalize the experience, enabling chatbots to provide members with 24/7 support that goes far beyond checking balances and making payments.
But that’s just the beginning. GenAI can also:
- Support better, data-driven human decision-making;
- Allow leaders to tap into collective intelligence, best practices and successful strategies to provide a better service for members; and
- Remove the mundane and make room for higher value work like building better customer experiences, imagining creative marketing campaigns and securing lifetime member relationships.
GenAI undoubtedly has the potential to transform credit unions in myriad ways, making them more efficient, enabling them to offer better member experiences and creating higher employee satisfaction through more valuable work. This potential can only be realized if credit unions embrace GenAI and successfully implement the technology, however, and successful implementation is not a given.
According to Rand, by some estimates, more than 80% of AI projects fail — twice the rate of failure for information technology projects that do not involve AI. Another research firm, Everest Group, estimates that 90% of proof of concept (POC) GenAI pilots may not move into production in the near future, or ever.
When embracing a new technology, the temptation is often to begin with the splashiest and brightest tool, regardless of whether it’s the right one for the need. This too often creates disappointing results, which can lead to abandoning the new technology altogether. But a more deliberate approach can yield positive results and long-term benefits.
Here are three tips to get started with GenAI implementation – and succeed.
Start with Strategy
- Be intentional with how AI supports larger business objectives, and member and employee experience strategies.
- Review your organizational goals and explore how, where, when or whether GenAI will enhance your strategy and accelerate your goals.
Work Out Where to Win
- Take the introduction of GenAI one step at a time. Focus on learning with simple AI integrations and build up to revolutionizing how high-value work is supported.
- Start implementing high-impact and lower-risk tasks such as process automation and build toward high-impact and high-risk tasks over time.
Pick the Right Partner
Choosing the right technology partner matters, especially as you move into the walk and run phases of your strategy. There are five critical criteria for picking a productive GenAI partner. Make sure they:
- Are an expert in banking regulatory compliance;
- Are an expert in banking workflows;
- Are focused on delivering outcomes, not selling products or solutions;
- Have a track record of delivering results to institutions like yours; and
- Understand your current and future appetite for GenAI and meet you where you are.
Let’s dig a little deeper. When choosing a partner and buying GenAI solutions, here are five questions to ask and what a vendor should be able to tell you, along with decision weighting for each.
1. Industry Knowledge and Expertise (30%)
What you need to ask: Is your GenAI solution specifically built and managed to meet the needs of banking?
What to listen for in their responses:
- Knowledge of regulatory compliance and adherence to all requirements;
- Adherence to general data privacy and security protocols (e.g., GDPR, CPPA and NIST 2.0);
- Ease of integration and requirements for integration with existing infrastructure (core and other banking-critical workflows/tools); and
- Previous experience working with similar institutions.
2. Model Transparency and Explainability (30%)
What you need to ask: How do you detect and mitigate bias in your GenAI models so we can ensure ethical and fair outcomes?
What to listen for in their responses:
- Banking industry data used to train models;
- Source of training data: Quality, veracity, availability and processing methods;
- Explanation of the model’s design, including algorithms, architecture and decision-making processes applied;
- Referenceability (e.g., citations to sources of information);
- Human oversight at critical points in essential workflows; and
- Approach to customer transparency and available recourse.
3. Transformative Potential (20%)
What you need to ask: Can we adapt your solution to our specific needs? If so, how does it scale for evolving requirements?
What to listen for in their responses:
- Ability to add your own data, policies and best practices to improve and enhance the model;
- Source of training data: Quality, veracity and availability;
- Human capacity and requirement to train and retrain the model; and
- Model explainability and interoperability.
2. Anticipated Gains (15%)
What you need to ask: What operational improvements and efficiency gains can we expect?
What to listen for in their responses:
- A clear performance measurement framework and reporting approach;
- Customer stories and case studies that report outcomes;
- Willingness to provide references from existing clients; and
- For innovative, new businesses, a willingness to consider risk or performance-based pricing.
3. Cultural Alignment (5%)
What you need to ask: Can you meet us where we are today and be a long-term partner in our GenAI strategy?
What to listen for in their responses:
- Crawl, walk and run optionality;
- Availability of strategy support from SMEs; and
- Cultural fit with the sales and delivery teams.
GenAI’s future role within financial institutions hinges on strategic integration with existing infrastructure and best practices. Our industry is regulated, we’re data rich and our teams are professionally skilled. But we all spend too much time on low-value tasks. The financial services industry is perfectly poised to embrace GenAI, provided it is the right sort of GenAI – the type that augments and enhances the work that credit union employees do, and the type that scales and strengthens collective intelligence.
It’s time to get started.
Corey Gross is Vice President and Head of AI Center of Excellence for the Austin, Texas-based financial experience company Q2.