As soon as generative AI hit the mainstream, hopes soared about its potential to save businesses time and money across various functions, including content creation.
But according to McKinsey’s State of AI report, adoption within financial services has been slow. In 2023, 24% of finance professionals reported regularly using generative AI tools at work. By 2024, this figure had only slightly increased to 26%, marking the smallest growth among all industries surveyed.
Perhaps this is due to what McKinsey refers to as “the most recognised and experienced risk of gen AI use”: inaccuracy, a critical issue in a highly regulated industry. And while AI is evolving almost daily, investment content creation is unlikely to be fully entrusted to LLMs anytime soon, if ever. That said, there are abundant opportunities for us to use AI tools strategically for specific tasks.
I’ll start by discussing how these tools can help you to understand your audience, since this is one of the first missions of a content strategist. In later articles, we’ll explore a range of additional key areas.
Audience insights are an essential element of any content strategy. Without an understanding of the people most likely to interact with your business, it is very difficult to know which topics and formats will appeal to them. Investment firms each have different needs and goals which will determine how thoroughly you decide to examine your audience.
You may have different groups of clients – for example, retail and institutional investors – in which case you’ll need to create and understand several segments. It’s also sometimes productive to create customer profiles or personas to represent different groups. It’s often useful to take a broad view of the investment industry’s audience as a whole, identifying where your brand’s specific audience falls within it and where your competitors are positioned.
One of the major challenges in understanding your audience is finding the right data. Typically, organisations in the investment industry suffer from one of two problems: either they have an overabundance of data and lack the resource to digest it, or they have insufficient data to work with. There are use cases for AI in either scenario.
Extracting audience insights typically begins with a substantial amount of research. We would look at the data available from website and social media analytics. Hopefully, we would have user testing, surveys, and customer interviews to review. We might supplement this with demographic data from sources such as the Pew Research Centre or the ONS.
Large language models (LLMs) such as ChatGPT or Claude excel at synthesising vast amounts of dispersed data that can be used to generate useful insights on audiences. AI-powered search engines, such as Perplexity.ai, focus on greater accuracy and credibility by using curated data sets and citing their sources, so they can be better for researching sophisticated, up-to-date topics quickly and at scale.
But besides concerns about factual or contextual accuracy, these tools can only give you answers based on training datasets and in some cases publicly available internet data – sources which may not be as accurate or useful as data your firm holds on its clients. For example, they may not provide particularly nuanced insights for one asset manager versus another.
AI can also be used to enhance your user testing. Research tools such as Maze and UserEvaluation can now digest audio or video inputs (e.g. user interviews or client service calls) and provide AI insights in minutes into your customer needs and pain points.
Maze has been used by financial clients including Vanquis Bank and Bpifrance to assist in their product design. Insights collected for this purpose can be just as valuable for content teams, so while this use case is less well-established, it certainly has potential. You would, of course, need to make sure that you have the necessary permissions to use data this way, and that it’s handled with appropriate security measures.
Audience segments are created by grouping individuals according to shared features. In the investment industry, this could mean splitting out clients from prospects, or grouping clients by age, portfolio value, or financial products held. Grouping based on these broad strokes is straightforward without AI, but AI tools can be more precise and are better than humans at making predictions.
Many customer data platforms (CPDs) and customer relationship management (CRM) tools now use AI to make predictions about a customer’s lifetime value or a prospect’s likelihood to convert. Similarly, they can create lookalike groups of prospects that share the largest number of traits with your highest-value clients. A couple of examples of these tools are Peak.ai and OptiGenie AI from Optimove. Predictions like these can help you to focus your marketing budget where the potential return on investment is highest.
Of course, these predictions are only as robust as the data they are based on. With insufficient or inaccurate data, AI tools will make ineffective decisions – and if you rely on these decisions, there’s a risk of creating a cycle of declining data quality.
Customer profiles or personas can help your marketing team relate to the audience and understand their motivations and goals. Of course, someone must first develop that deep understanding of the audience independently in order to create the persona. What is their financial literacy level? How high is their risk tolerance? What investment news do they follow? Alternatively, you can use an AI tool to answer these questions.
Various tools exist specifically to create personas, including Delve AI and InstantPersonas. They work by using the data you have available from, for example, Google Analytics, Meta Analytics, or data from Salesforce or another CRM, and combining it with public data e.g. news, reviews and forums.
Naturally, these tools work much faster than a person can. Plus, since AI analyses data so thoroughly, it may pick up details that a person would overlook. However, some might argue that an AI-generated persona can be overfitted to reflect certain statistics or data points. If the role of a persona is to breathe life into faceless statistics, perhaps a human has something ineffable to add.
Some persona tools offer the option to create personas not just for your brand but for your competitors, illuminating the similarities and differences between your clients and theirs. Though it may be difficult to validate the accuracy of the output, this could assist with identifying what’s unique about your brand or which markets you’re failing to capture.
AI can also help with more traditional methods of competitor analysis. There is a wide range of service providers that scan public platforms for mentions of your brand and its rivals, or other relevant topics, and use AI to present actionable insights. Tools like Crayon, GapScout, or Amplyfi might be very useful for resource-pressed marketing teams looking for keener insights.
It remains a strategist’s role to determine where your business is trying to get to and how. AI may be able to help spot opportunities, decipher vast amounts of data, achieve efficiencies and provide insights. But as with content itself, strategy and differentiation will continue to depend on experience and perception. As we will show in our following pieces, AI can provide enormous advantages, but humans should decide how to apply them.
If your business is looking to better understand its audience and enhance its content marketing – and whether you’d like to use AI tools or simply human intelligence – Copylab can help.
Get in touch to find out more about our content strategy and other services.
