Getting the edge with AI in financial services marketing

For many years, AI has been helping streamline all kinds of workflows to improve efficiency. From data entry to data analysis, AI can have at least some impact on every step of the process.

However, given the launch of ChatGPT back in late 2022, the marketing side of financial services has been hit with a wave of new opportunities (and pitfalls) to be aware of. Promising the same sorts of efficiency increases seen in other business departments, AI, through the use of large language models, is a hot topic right now and will continue to be so for some time.

While AI does indeed carry benefits for marketing in the financial services sector, there are some things to be aware of before diving in. Here are some of the best ways of utilising AI in your marketing output.

Ways of Using AI in Financial Services Marketing

Content Ideation

Content ideation is, by far, the task most reliably assisted by AI.

Coming up with content is hard. Not only must you decipher interesting angles on subjects, but such angles must be delivered in a way that your audience cares about.

If you’re facing a spot of writer’s block, or creative dissonance, using a chatbot can be a great way to talk through your ideas while coming up with new ones. Simply provide a short description for your situation, followed by what it is you’re looking for.

The best prompts for content ideation typically include some information about your audience, your brand’s tone of voice, as well as some examples of content you’ve written in the past. Providing that past content is crucial, as it’ll help prevent the chatbot from treading over past ideas, saving you a lot of time in back and forth.

Given the amount of regulations and nuances within the financial services sector, you’ll need more than just working titles to get the most out of your content.

Once you have your ideas, you can then ask the chatbot to help you flesh things out into presentable writer’s briefs. In our experience, these briefs can get quite repetitive, so we do advise going in afterwards to tweak things to make them less formulaic.

Text Analysis

To be clear, we never rely solely on AI for written content. The chances of inaccuracy are just too high to warrant such a risk, even given how quickly results can be delivered.

However, we cannot ignore the text processing capabilities of modern large language models. They’re called language models for a reason, after all.

So, while we don’t trust LLMs to write content from the ground up, we are more than happy to feed it content that has been previously written, while asking it to proofread for usual grammatical mistakes, and even make suggestions for improvement.

Something that ChatGPT shines at in particular is sentiment. You can give it a small sample of content and ask it to “describe the sentiment of this text while providing evidence”.

You’ll then be met with a reflection on how your writing is perceived by the chatbot, with it listing off positive, negative, or neutral in its feedback. This can help critique customer reviews en-mass or compare the sentiment of your competitor’s web pages to yours, while using the findings to refine your messages and make them more aligned with what your customer personas expect.

Regulatory Guidance

AI is a touchy subject in the eyes of compliance and regulation. Trusting a language model to produce written material is shaky because it isn’t as well versed in the regulations as a qualified expert, but the efficiency benefits are hard to turn down.

There is the possibility of feeding regulatory guidelines to a language model to train it. ChatGPT is quite useful for this, specifically under its relatively new “custom instructions” feature.

Enabling custom instructions allows users to enter a series of predefined parameters for all conversations to be based on. This is the perfect place to provide some quick training around the latest rules, like consumer duty for example.

It’s not perfect, and human inspection is still always advised, but it could save you some time in the long run.

Things to Bear in Mind

While AI does present many opportunities, it isn’t without its concerns:

Google’s Helpful Content Update

Back in August 2022, Google rolled out a new update to its search algorithms dubbed the  “Helpful Content Update” (HCU). The stated purpose of this update was to ensure the web isn’t full of blog articles and web pages that are written solely to gamify search results.

Reading between the lines, it’s possible that Google anticipated the incumbent launch of more robust large language models, and instilled this update to try and circumvent the influx of newly AI written content landing in its search results.

This of course is just speculation, but the HCU certainly should make a user think twice before publishing blatantly AI written material.

The HCU was recently updated, and now directly references non-human written material, stating that it’s absolutely fine to use such content on your website; provided that the material is legible, human-focused, and accurate.

EEAT principles

Speaking of human-focused, high-quality website content.

Back in late 2022, just after the aforementioned helpful content update was rolled out, Google added another E to its EAT (experience, authoritativeness, trustworthiness) guidelines for web content.

This newest E, meaning experience, is all about showcasing an author’s lived-experience with a subject; something that can’t be explained with AI.

You can learn more about EEAT in financial services, here: What is EEAT and why does it matter in financial services?

AI deterioration

A paper from Stanford University has come to light in recent months, highlighting that, since its release, ChatGPT has gotten progressively worse at queries compared to when it was first launched.

Particularly around arithmetic, ChatGPT had started to inaccurately answer posed questions, while still giving sound reasons for its work.

Need help leveraging AI? Speak to Balance

At Balance, we’ve adopted AI into our internal processes to make our team as effective as possible.

Language models have made us faster and more efficient at creating quality material for our clients. Want to help your team do the same?

Don’t get left behind in securing your edge using AI. Contact us for in-depth AI training delivered by experienced financial services marketers and we can create your session at a time to suit you.