The 3-step Roadmap for Pursuing AI Projects

Outlining the benefits of AI is one thing. Integrating the tech helper as an extra puzzle piece for your fund marketing is another. And there’s no better time to get started on your next project.

 This is, in part, due to its proven capabilities, which are yet to grow in scope. But it’s also due to financial institutions’ investment in what AI can do. Nowadays, some 30-40% of the marketing budget is being allocated to AI technology.

So while a marketers’ AI experimentation could be considered a passion project, trial-and-error can yield great effects on asset raising and ROI to please those allocating that cash, whether it’s for building a virtual assistant, or implementing a sentiment checker or GenerativeAI language tool for content creation.

Like performing persona research or building out a monthly content calendar, an AI roadmap  is needed to get the most out of it specific to your project’s needs, which can be simply defined in three stages.

Laying the foundation: identify your goals and use cases

Jotting down intended solutions where AI may be able to improve existing strategic outcomes, mapped against prior methods and results, is key to validate the initial worth for any new AI-backed marketing project.

According to most fund marketers, these are the common pain point areas:

Goal: Greater lead generation

By pinpointing where data processing and automation capabilities can supercharge individual marketing missions, it lays the foundations for where AI may be able to fit in comfortably.

Implementation: planning pilot projects

With initial ideas drawn up, AI utilisation still needs to be planned for, road tested, then evaluated. Different institutions may require very bespoke uses for AI too. But so long as it can provide a high impact on ROI, it is worth focusing some time experimenting with different AI tools to assess the difference they make.

To illustrate such a specific AI-backed use case, we trialled our own developed chatbot – DasBot – using Athena AI, similarly useful for an asset manager’s website to handle investor FAQs. We wanted to see if it could replace a simple ‘Help’ feature using data processing.

Instrumental to an AI project is making sure that your data is well prepared before being integrated by the technology. An AI tool can function to full capacity when data is clean, organised and made readily accessible. In this case, we used Athena to assimilate twenty years of ProFundCom content data, armed with knowledge to answer questions we had around marketing best practice.

In response to the user asking for “a good LinkedIn strategy”, DasBot returned a neat list of actionable ways to boost engagement on the social media platform based on our historical insights. With different search criteria, this could equally show a fund marketer new techniques to improve their SEO, or to build email templates. The resulting information given could even form a basic structural outline for a blog, webinar, or white paper.

Scaling and optimisation

By using this AI as a crash test dummy internally, it helped identify user experience niggles that may need ironing out, for instance, if the supporting information was proficient and correct or if the navigation was simple.

This leads us to the final stage of an AI project: assessing a tool’s effectiveness before deploying it to create external-facing content, or as a product fit for customer use.

• Is the AI easy to use?
• Does it accomplish the goal it sets out to achieve?
• Can it be improved (through data or design)?
• Will it fit in with investors’ marketing preferences?

This last point is a key consideration. Even after developing any AI tool to suit a brand’s image, will it appeal to, or distance, existing clients? Some investors may feel more comfortable emailing or calling rather than using a chatbot. Identifying an AI tool’s popularity against past interaction data with other software could help to refine an approach that achieves a healthy balance.

Only through optimising AI prompts and outcomes can better results be achieved. If you’re looking to develop and scale an AI tool, the following considerations also need to be taken into account for the heavily regulated fund world:

Compliance. Is your AI able to handle sensitive user data in line with regulations to avoid reputational risk?

Team training and support. Can the AI tool be rolled out internally? Adequate training, or running AI through a centralised, cloud-based platform can help, which also creates an authenticated audit trail.

Ethical concerns. AI can be prone to algorithmic bias: repeated (often unintentional) errors or hallucinations based on who trained the model, prone to manipulation and poor misrepresentations of some users over others.

Plagiarism. Your own data and content is a walled garden! Training models with data outside your organisation could infringe licensing and copyright laws. Copying and pasting copy straight from a GenAI platform is also plagiarism. Tools including ChatGPT or Jasper should provide a structure for a marketer to fill out with their own expertise. After all, as fund managers’ attitudes to risk or opinions on investment strategies or portfolios are so varied, views generated by computers cannot match this idiosyncratic human thought in quite the same way.

While these concerns remain, with greater AI adoption in financial services set to rise, marketers’ ability to take their AI tools from ideas to practical and safe execution is also growing. Whether slotting in AI as an automated backbone for marketing operations, to sense-check your content pieces, enhance your engagement strategy, or empower your investors, the possibilities are endless. Get experimenting today!