Episode 55 - The Do's & Don'ts of Using AI for Your Financial Institution Content

Hi, there. I'm Meredith Olmstead, CEO and founder of FI GROW Solutions and we are an inbound marketing and sales consulting agency, and we work exclusively with banks and credit unions. And I am here with Kristin Mock. Kristin is one of our inbound marketing strategists and we were just having a really awesome conversation about AI, artificial intelligence. So Kristin, say hi.

Hi.

So I'm learning a lot about this technology obviously, as we all are, as it's coming on really strong over the last couple of years. And Kristin and I were just having an interesting conversation about how to better leverage the technology of AI for marketing purposes, more specifically around content, right? Because content is king online for banks and credit unions, that is accurate. And so what we are trying to figure out as a team internally and for all of our clients is how can we use all these new AI tools that are popping up on the internet and out there to better create and manage and leverage content for clients and for all of these institutions, these financial institutions? So I said, let's push record and talk about this a little bit more and share this with everybody.

Generally, most of us are probably already using AI if we're working in marketing for a bank or credit union and we may not even realize it. So Google Ads, predictive text when we're typing, Grammarly, all of those kinds of things, those little tools that we have in the background, are already utilizing AI. So we know that that's happening. But what Kristin was trying to help me see a little bit more clearly is what are the dos, what are the things you should be doing with AI and what are the don'ts, what are the things that you really need to avoid? So let's unpack this a little bit. So Kristin, first thing, when you think of AI and you're maybe talking with a client or talking with me or somebody out there who's a marketer at a credit union or a bank, what are you going to tell them? What is the first main thing that they should be thinking about how to leverage that technology correctly?

I think the biggest do is generating ideas. I think that getting a list of topics or a list of blog ideas is a huge do. When most people are trying to use it to just write that content for them I think that's the biggest don't because they're just jumping in and trying to say, "Write this blog for me." And it's ... Just use it for your ideas. Use it for your list. I think that's the biggest first do.

So you can put in a question like, "What should I be creating content around a product or service line?" Or, "What are people asking? What are problems that people are trying to solve around applying for a mortgage or opening a checking account online?" or whatever. And AI can generate a bunch of ideas for topics, for blogs, for articles, for questions and answers, that kind of thing. Okay, cool. What about keywords? Because I know we do a ton for clients around research, keyword research specifically, and really making sure that our content or the content we build for clients is very strategic and is created and optimized to show up on search over time.

There's really two ways to handle keyword research. There are some tools with AI that will do some of that research for you. Yoast is a good example of that. But if you're already using a tool that is comprehensive and that you enjoy, you can take that keyword research and plug it into an AI tool and it can help build an outline for whatever it is you're writing, a webpage or a blog. And it will help incorporate those words and structure them for you.

Okay. So if you're going to send ... Say you're creating a new ebook, checklist, blog article, and you want to send it to writers, but you want to give some bullet points for what you want them to cover, that could be a good starting point and a use for a tool? Okay. Now the other thing that really blew my mind when we were talking about this was you said that you can actually use AI to analyze customer data around content, reviews, feedback. How?

Well, so a lot of your analysis tools are already using some AI to show some predictions but a great way that you can use it outside of some of your analytics tools is for reviews. So if you're getting reviews on Google or Facebook, you can download those and you can paste them into an AI tool and ask it to draw conclusions, to find trends. The default is to find positive trends that will say, "Your customer service is really excellent based on these reviews." But you can specifically call out and say, what are the negatives? And it will find and bullet point for you, these are your areas of opportunity.

Oh, wow, I wonder ... So I imagine you could probably do that if, say, you ran a member survey. So if you get back like a list of a hundred open-ended answers to tell me more on a survey and you don't want to spend three hours reading every single one of them, you could put them all into an AI tool and get overall trends. That's really cool. That could be a very big time saver but also still pull out and glean the main points that needed to be taken away. I like that a lot. What about, okay, you also said in terms of editing or fact checking content. So how do you use AI for that?

Yeah, so one big flaw in AI right now is that it might say things or write things for you that are not always factual. So one thing that you always need to do is double check facts. You still need to be that expert on your product, on your service, to make sure that if AI were to say that you're offering 0% APR, that might not be true. It might pull either 0% from one place, the APR from somewhere else. It's just something that you're always going to have to fact check everything that comes out of AI because it is pulling it from different places. And then for editing, yeah, editing is huge. You can do editing two ways. You can do it as you write. There's things like Grammarly plugins or ProWritingAid has a plugin that's very similar to Grammarly and it will add commas and fix misspellings and any grammar mistakes that you're making as you write.

But you can also plug your content in to some of these editors and ask it to edit. You can use the Hemingway Editor is one that I use a lot. I tend to use jargon that maybe is not going to appeal to the audience or use a lot of commas and end up with really long sentences. So the Hemingway Editor is specifically for simplification, to simplify out your sentences, your jargon, and just make it a very straightforward sentence. So that one has a very specific use case. But some of the other editors, you can also ask it, "Where can I expand?" So not just your grammar edits, not just your misspellings, but to say, "Hey, I have a 300 word blog article and I need to make it 800 words. Where can I expand on this?" And it'll pull out those topics that might have more information that you could include.

Wow, very cool. Okay, that's a great way. A lot of times it's just that writer's block, right? When you're working with something or you're trying to push in a different direction and you can't quite get your head around it or think of anything in the moment, it could be potentially a really great way to overcome those little sticking points for people. I love it. All right, so those are a lot of really great dos. What about the don'ts? Because there's got to be some, don't do this, be careful kind of stuff.

So I mentioned earlier writing an entire webpage or blog post because I think that that's the number one way that marketers are expecting AI to work. The first thing they do is they say, "Write an article about credit cards." Well, it's not going to do that for you. That is the very first, most important don't, is don't just expect it to do your work for you and write an entire webpage or entire blog post and it to be accurate and entertaining and valuable to your audience. Because it won't.

Got you. All right. So it's not going to replace you. I also love this. We were talking about one of the things that we work with our clients on is trying to become trusted advisors and thought leaders in their markets around their products and services. And most of them obviously are financial products and services. So how can AI hurt that goal? We want to be careful, right? We want to make sure that people are kind of standing out in the crowd. So how do you have to be careful with AI?

Right. So the way that AI works, at least right now, is that it is crowdsourcing information from the internet. It is taking the things that are the most talked about in the way that they're the most talked about, and it is just gathering those items and churning them out, which is exactly the opposite of what you want. If you want to be a thought leader, if you want to be unique in your space, AI is not going to do that for you. That's where you have to come in. It might help you come up with some ideas but you have to understand that the things coming out of AI are the things that other people have already put on the internet. So you are automatically going to be not unique by using that.

Got you. So it's not really super strategic. I mean, it can generate some ideas, but it may not be able to tell you the best way to utilize those ideas, those topics, that content, and it's really not going to draw conclusions or help you make really, really thoughtful decisions or move in a certain specific direction. So that is logical. I mean, I think that makes sense. What about the SEO issues with AI created content?

Yeah, we talked about using AI for SEO keywords, which is useful, but if you're just expecting it to write a sentence or write a heading that is going to rank on SEO, that's where it gets a little tricky. Because like I just mentioned, it's using things that are already commonly talked about on the internet. So if you're pulling these words and these sentences that are frequently used on other websites, then that's just giving you a really high difficulty. And then it's not going to be very easy to rank SEO. So for now, at least the way that AI is working, I would say that do not expect anything that it comes ... that it churns out for you to just be ranking well on SEO. That's definitely something you have to do yourself.

Yeah. That said, I know that we did chat about this in a team meeting that right now Google is not dinging artificial intelligence, like AI assisted content. So it still will rank, it will be crawled. It's not going to be penalized by Google yet, but potentially I guess that could also happen. So we also want ... So what about brand voice? That's really important for most businesses, obviously, but for financial institutions for sure.

Yeah. So that one came up with a client this week actually, is that they were really disappointed that the AI tools they were using were not successfully writing in their brand voice. So that's definitely ... You cannot expect it to just know your voice. Even if it's gleaning content from only your website, it is not going to nail that. You can add prompts. When you're asking AI something, you can add into those prompts to write it in someone's voice. You can try any celebrity, any author, sometimes it provides a great laugh. Just say, "Write it in this person's voice." And sometimes it really nails it but your brand voice is going to be a little different than a celebrity. So you can try adding, "Write this in a friendly way, in a casual way. Write this as a helpful business." But you are still going to be tweaking a lot of that to make sure that the wording and the sentence structure matches your brand voice.

Got you. So I think the biggest takeaway is that you can't skip the review process. You're never going to get content from an AI tool that you can just copy, paste onto your website or into an email, not review it, not edit it at all, and hit send or publish. That's just probably not going to happen. I mean, so I guess that's good news in that we're not all about to lose our jobs, right?

Yeah, no, I mean, we have a long way to go before AI is going to be just doing our jobs for us, use it as a tool, but edit away.

Yeah. Awesome. But it sounds really, really useful and a lot of potential for making things more efficient and helping to expand in terms of topics and directions for where you could consider going with content. So I love it. Love these ideas. Awesome. Well, thank you so much, Kristin. I really appreciate you sharing your expertise. You are the content expert on this stuff. And so if you all are interested in learning more about marketing and sales for financial institutions, please visit us at figrow.com. We have lots of other podcasts, blogs, ebooks, case studies, all that kind of stuff. Lots of helpful resources there. But otherwise, in the meantime, let's just all get out there and make it happen.