One way of defining marketing is as a tool to demonstrate value to your target audience. What Artificial Intelligence offers is the ability to transform what you can demonstrate, by powering the insight you can offer with the muscle of computing power that AI harnesses.
There is huge opportunity in financial services data – from both publically available and premium data streams as well as what your entity accumulates – for AI to enable new insights for a discerning audience to consume through both content and other marketing efforts, like press events or symposiums.
Automation makes things easier and AI automates repetitive decisions so you can make harder decisions, instead. Switch out the word “decisions” for “observations” and therein lies the value of AI to financial services marketers: offer observations that were never viable before.
Positioning for AI works for content production, too, and the latter will benefit from the change immediately. Focus on simple rules based around repeatable human action that later you can parallel with AI: make it easy to source and sort information.
1. Establish the authority and protect it at all costs
You need a benevolent dictator, respected in their understanding and track record in representing financial services with authority and capable of delivering what the target audience is interested in. That person, or persons, represents a surety all departments can have faith in.
That does not mean other opinions do not have an audience (see 2), but editing-by-committee will kill any initiative. What scares organisations is risk: How do you discuss publicly topics that are confidential or sensitive and around which disclosures are carefully monitored and subject at all times to scrutiny? How do you turn AI loose, safely?
Establishing a process with respect to a few key principles is enough and handling risk early is easy: draw a line between disclosable information and opinion based on secret stuff at the start and have inputs submitted clearly flagged. The process itself negates risk and makes outputs clearly traceable back to source. Just make sure you have the appropriate consent from all stakeholders and the legal cover in how you present it.
Input from across the spectrum is both controlled and unleashed by empowering a clear hierarchy. You then have someone with overall visibility that can make the final call on what to publish, which helps calm any nerves. Newsrooms deal with sensitive information every day – a clear hierarchy makes it possible.
2. Engage everybody in the process
Making everybody feel heard cultivates buy-in and improves your output.
Have a process in place to get input from everybody on quantitative analyses, assertions, raw data, or even just a market or news event. There are endless ways to do this, so focus on making it an easy process that encourages familiarity. This will help in engagement early on and in driving the quality of the work long into the future – and it means you can quickly deploy to gather answers when Black Swans strike.
When I say everybody, I mean EVERYBODY: there is no cost to ensuring you get input from the most junior to the most senior people quickly, simply, efficiently. This means organisational buy-in, which is critical to both effective marketing in complex industries and innovating in AI.
This also makes answers better. The wisdom of the crowds is powerful and acts as a safeguard against manipulations (and accusations thereof).
More people involved means better data, marketing, and engagement. The best-marketed entities in finance and consulting all receive mandatory contributions from the workforce.
3. Brevity and speed control input, control output
Ask people and departments to submit precisely how you want them to and let that build out a process that ends in an output you can subject to dissociated decision-making. Control what they give you by restricting the time and length they have to answer.
Establish how to separate data and observations, then tabulate, combine, and advance. Formulaic and repeatable answer formats are adaptable and easy to feed into models and, later, deeptech AI stacks that take things to further. They also give human intelligence and creativity within your organisation a stage to shine on and help people bond around the task.
Always gather both a quantitative and qualitative answer for everything. Separate their input forms into observation (“rate these things from 1-10 on X”, “what do you see here” etc), meaning (“rate these causes”, what do you think caused this”), and prediction (“how would you rate the likelihood of these outcomes”, “what are the likely/unlikely outcomes”). Force speed and brevity – don’t give them too much room to answer.
Ambitious routes to using AI open themselves up naturally. Once you have process data, set benchmarks for interpreting and predicting answers: Imagine refining one conclusion down to 100 yes/no questions, then determine how many or which of those would guarantee that conclusion.
Brevity, speed, and the qualitative match to every quantitative value becomes more rewarding the further you go. That your process is based on answers from qualified people, and conclusions are traceable back to source, also keeps you and everything you say compliant.
4. Start with the process, not the tech
Enforcing brevity and speed at the start covers you into the future because it prioritises process: process is what matters and you fit technology around it.
You can and should use off-the-shelf tools to simplify tasks. Once you nail a process down you can start to think about how you might stitch parts of that process together into an automation. If you can do it simply or through a series of very simple steps that is even better: easier to perform, understand, and scale out. The magic of AI is not in general purpose functions, but in the specialised processes they enable people to create. The best always have a simplicity in how they solve a problem, like good writing or engineering. Whether your tech is better than someone else’s is irrelevant – use what moves you forward and lets you focus on the next problem. Brilliance lies in the sequence you build.
5. Observations are more important, not less
Get information, tabulate it, automate that, automate the next, plugging in tools when necessary. Make fast decisions on the output enough that you can create a formula to tabulate those. Hey presto – the beginnings of your own AI, even if (especially if) it is all stuff you can technically do in Excel.
But data and AI-powered analysis is not going to be enough on its own – it has to come with interpretation that pushes things forward. In finance everybody sees the same markets, economies, events, and global activity that informs everything else. What makes organisations special is what they do with what we all see – what makes AI special is what it makes possible to show.
A lawyer friend asked me recently if he should publish anything online. My answer: “the law is the same regardless and if you’re smart you can guess the basic answer, but what you write offers a glimpse of working with you. This is a chance to show people the value of working with you.”
