This article is the final in a three-part series exploring the watershed moment that AI brings, the journey to be compliant, and the AI use cases within a credit union.
AI is no longer rare in the credit union movement. Many credit unions already employ AI in their call center, their lending origination system (LOS), and knowledge management. This article is about going beyond the vendor-packaged uses of AI to show leaders how to use it as a strategic tool, consultant, and sounding board for your most radical ideas. Using AI in this way is something you can do today with AI tools like ChatGPT, Gemini, Claude, and others. You do not need a vendor to do anything, just start your AI tool subscription and use the AI prompt framework explained in this article.
We’re all struggling for time and AI creates time for executive leaders. It does so by getting fractional FTE efficiencies in nearly all executive functions of planning, modeling, and policy. Using AI in this way is scalable for all roles in the credit union, but for purposes of this article, we will focus on executive strategic thinking, data, and planning. Let’s begin by unlocking how executive leaders can use AI for strategic impact with the PCACI Method.
Super strategy with AI, using the PCACI Method
The PCACI Method is a supercharged way to use AI to free you of the burden of drudgery so you can focus on taking care of your membership. The mnemonic device stands for Persona, Context, Ask, Constraints, Interview and is a great way to interact with AI for powerful insights and outcomes. This is so you can make meaningful changes in your credit union to keep pace with a market that is violently re-making itself in the new AI economy.
The PCACI Method is a way to structure your prompt to the AI for move-the-needle insights to keep your credit union competitive. I used to roll my eyes when people would say they were a “prompt engineer,” but over the last few years, I’ve seen the difference that a well-crafted prompt can make. The PCACI Method is a framework for prompt engineering to get work with AI as a powerful strategic partner on your team.
Without getting into the technical aspects of AI, it is important to know that modern AI providers trained their AI on nearly the whole body of human knowledge. At first glance this seems like a wonderful thing—and it is—but for credit union purposes it is not ideal. I liken it to asking a professor that knows everything and gives you very general answers when you ask specific questions. How do we talk to the 30-year expert of community banking that knows the NCUA rules and regulations inside and out and not the general knowledge professor?
This is where Persona in PCACI Method comes in. You need to tell the AI what persona to assume, this sharpens the AI’s generated content to go from intriguing to revising the strategic plan. We all had AI tell us that our ideas are smart and interesting, but that is not what you need to hear when you are deciding on a five-year strategic course to recommend to the Board. With Persona, you can dramatically shape how the AI thinks and responds. So how would this look?
Persona
You are a 30 year expert in the credit union industry with extensive knowledge in NCUA rules and regulation, as well as GLBA. You have been a credit union CEO and board member. You are my strategic consultant who is no nonsense and will call out bad or weak ideas quickly. You want my credit union to succeed and are not afraid of hurting feelings. You are patient and will talk through things step-by-step and explain them to me as an executive coach.
Now, let’s move on to Context. Context is where you provide the AI with all the information you have that you feel might be useful. This can be content you write in the prompt window, but you can also refer to other websites, government published statistics, or even uploaded documents. When I am working with credit union leaders, I encourage them to upload their redacted strategic plan or a previous one if they are worried about proprietary data leakage. Context is important in helping the AI give you a tailored response that is hyper-relevant to your credit union. My credit union is in Baltimore, so I might structure the Context prompt like this:
Context
We are a credit union located in Baltimore interested in the population growth models for the next five years. Use current federal and state census data and projections in your response. Use NCUA historic credit union call report data for my peers for the past five years. Uploaded is the five year strategic plan as part of your context.
The next part of the PCACI Method is Ask, which is where most AI users jump straight to and are predictably underwhelmed by the results. AI will generate a response for your prompts, but the PCACI Method will refine it into the credit union-growing impact you’re looking for. So how does an Ask look in the PCACI Method, here’s one that I used to challenge and improve a strategic plan:
Ask
Analyze the uploaded strategic plan and compare it against the relevant government data. In your analysis, consider the key metrics like member growth, deposit growth, and non-interest income growth. Ensure the strategic plan has sufficient tactics to achieve the key metrics in the five year time horizon. Identify gaps or issues in the strategic plan and make recommendations on how to address those gaps and issues. Consider the talent of the management and identify where there might be talent gaps and provide recommendations. Provide a potential Q&A that the board might ask of my management team so I can be prepared for these questions when I present the revised five-year strategic plan to the board at our next strategic retreat.
The next two parts of the PCACI Method are optional, but I found they are very helpful for important matters. Generally, I will use Constraints and Interview if it is a matter of strategic importance, those things you only get to do once or affect people’s lives. Constraints are simply the instructions to the AI on what not to do or how to structure the response. This is how I might structure a Constraints portion:
Constraints
Be direct and succinct in your responses. Do not be sycophantic, tell me the realities in your response as this is important and impacts people’s financial lives. Do not apologize. Structure your response in clear sections related to the strategic plan with citations, identifying the issues or gaps, and offer 3-5 recommendations on how to address them. Be sure to cite government data in your response. If you don’t know, say so.
At this point, you would think we have everything covered, but I found that the Interview portion is the key to ensuring you have. Rarely in my prompt preparation have I actually covered everything the AI needs to generate a thorough and impactful response. If I did not give it a chance to ask interview questions, it would have responded without the information, making the response weak. Interview allows the AI to consider what you’ve told it and ask questions to clarify things. Interview looks like this:
Interview
Interview me one question at a time, to clarify anything you need or would like more information about before generating your response. Do not ask more than ten questions. Prioritize your interview questions.
After entering in all the PCACI portions you are ready to hit enter on your AI tool of choice. Often AI will ask you several follow up questions, honoring your Interview portion of the prompt. Just respond to its questions as best you can. If you don’t know, it’s OK. Often AI will identify the missing knowledge as homework. You don’t have to limit AI in the interview questions. Though, I found it helpful to make AI more economical in the interview questions.
Frequently, people report that AI will offer at least a few high-value recommendations they had not previously considered, which is where the true value lies. After all, isn't that what a good consultant does for you? A good consultant should not just agree with your plan, they will spot the good, recommend fixes for areas that are weak, and offer ideas you have not considered yet. AI is the expert at your fingertips, but without the invoice.
In this example of using the PCACI Method, we explored how to use this expert at your fingertips for strategy. But what other experts can you call on? In my experience, I found that AI can be a good regulatory expert and a good marketing expert to name a couple. Just remember to change that expert you need in the persona. Asking a regulatory expert on the best account acquisition strategy for your market isn’t likely to yield a good outcome.
Next steps
We all know that banking is a highly regulated environment. Likely, there are credit union leaders reading this article who are pulling their hair out over regulatory and GLBA concerns. In my previous article, I covered the compliance roadmap for credit unions, so we’ll not cover that again. However, it should be noted that the example above does not expose any sensitive member information, assuming none is in your strategic plan. It does expose potentially credit union sensitive plans. Consider using this strawman strategic plan for ABC Federal Credit Union while you get the hang of the PCACI Method for strategic planning.
You can use the PCACI Method for other conversations with AI. There really is no limit to what you can use this method for. I recently used it to help me create an engagement plan with my overseas developers and settling an insurance claim. This method allows a creative credit union leader to play what-if scenarios and identify how to improve operations in their credit union. We have reached the point where AI will be everywhere, it is now about AI orchestration.
As a credit union leader, consider the value and cost of your executive team. Imagine what it could be if you got only 10% more return on investment from each teammate. Using the PCACI Method promises to deliver that and much more. Establish management controls and encourage your executive team to use AI. It is not taboo. Rather it is now necessary for credit unions to stay competitive as the market is remade in this AI economy. Now is the time where organizations learn to use AI or the market will quietly help them exit. AI unlocks scale never before possible, but we must not forgo our humanity, for it is not technology that cares, but our people.
