AI and data: A perfect match?

AI has quickly become part of an adviser’s engine room, but to use it effectively, you need high quality data. A panel of advice experts recently shared how to combine the two to achieve stronger client outcomes.

Artificial intelligence (AI) is changing advisers’ workflows, practices and even ways of thinking at a rapid speed, but it’s becoming increasingly apparent that AI’s output is only as good as the data it’s fed. Understanding this will be critical for the profession as the appetite for AI grows.

Research conducted by the Financial Planning Standards Board (FPSB) this year found the profession is at an inflection point, with almost four in five advisers believing AI will improve how they serve clients.

The global survey also found one in three advisers use AI for data collection, highlighting advisers’ growing reliance on AI to build client profiles.

“We are witnessing a pivotal moment in the financial planning profession as financial planners embrace AI to work smarter, allowing more time to engage in deeper human connection with clients such as navigating difficult conversations that impact financial decision-making and providing clarity and support to stay on track to achieve their life goals,” the FPSB’s CEO Dante De Gori said.

At the same time, there’s recognition that AI isn’t a complete solution and has its challenges: it’s only as strong as the data advisers provide, some applications raise security concerns, and in some cases AI can generate inaccurate insights.

With this in mind, the key question becomes: how can advisers best apply AI to understand and serve clients more efficiently? High-quality data is a critical part of the answer.

At ensombl’s recent All Licensee PD Day, three experts – intelliflo’s Stu Alsop, Guidance Financial Services’ Paul Benson, and Coastal Advice Group’s Mitch Ramsbotham – shared how advisers can combine AI and quality data to transform their businesses. Here’s what they said.

Quality data underpins everything AI can do

While AI is a powerful supplement, without the basics nailed down, the reality is it can’t perform in the way practices need it to. Good quality data has to sit beneath any AI process. In other words, practices first have to refine their data before they can input it into AI to inform the way they practice. While it doesn’t necessarily have to be a completely manual task, it does have to happen before any practice can expect optimal results from AI.

“Quality data is really the bedrock for anything you want to do in business,” intelliflo’s Alsop said at the All Licensee PD Day.

For Benson’s practice, the next question will be how to apply AI to fuel the support side, which again, is underpinned by quality data.

“It’s just evolving so quickly… The tools and the capacity of what it can do – you’re thinking of new applications all the time – but it does need a new data source in which to extract the information in the first place,” Benson told the All Licensee PD Day.

Conversely, relying on poor-quality data can put a business at risk. For example, if AI builds a client profile based on incorrect information, the resulting insights could undermine trust and damage the adviser-client relationship. It could also lead to reputational damage for the practice itself.

Best practice involves customising AI

We’re all familiar with popular AI applications developed by global companies, but the key for practices is to make AI fit into what they do, as opposed to fitting into cookie-cutter AI models.

There’s a common misperception that such customisation comes at higher cost, but that’s not necessarily the case. Through scalability, trusted AI applications can be integrated into advisers’ workflows in a way that aligns with their ways of thinking and doing things. Increasingly, we’re seeing practices build integrations through partnerships to better serve their clients. This will be a huge space to watch.

“[intelliflo] has had an open APIs for 10 plus years in the UK and since when we launched in Australia in March 2023… allowing people to not only use off the shelf integrations, but also use custom integrations with the software and systems they’re using today,” Alsop said.

In the next six to 12 months, we expect AI to develop and perform in ways the market hasn’t yet fully grasped. Despite that, the key will always be having quality data to input.