It is a fact not universally acknowledged among advisers that – regulatory speaking – there are only three things they need to include in a suitability report.
What to include in a suitability report
The first is the client’s demands and needs. The second is why the recommendation is suitable based on that information. The third is the possible downsides. For years, debate has raged among advisers over the best way to document all that. How short can you make it? What’s the best layout? How many tables and charts do you need?
A whole host of technology firms and software solutions have sprung up to try and provide answers. But no technology has made quite as many waves in recent months as ChatGPT, the AI text generator that can do everything from answer factual questions about historical figures to write your next university essay.
But can it write you a suitability report? I decided to find out.
I asked ChatGPT to do the following:
“Write an FCA-compliant suitability report for a financial adviser with a client that has an income of £250,000, alongside £1m in liquid investments, £450,000 in a personal pension, expenditure on annual school fees of £20,000, and outstanding mortgage payments of £100,000.”
This is obviously a very much made up scenario. It may not be your typical client, but I hope it reflects a potential client at least.
What did ChatGPT write?
Let’s see how ChatGPT did shall we? After a brief introduction laying out the client’s financial position as per my prompt, the auto-suitability report lists their objectives as the following:
- To achieve their goals with a low to medium risk strategy.
- To ensure that their retirement income is sufficient to maintain their current lifestyle.
- To secure their children’s educational expenses.
- To pay off their mortgage within the next ten years.
Does that meet the first requirement of the FCA: to document the clients’ needs and demands? Well, arguably not, because you’ve not actually asked my client about what their goals are, where they see themselves, and how they would personally prefer to get there.
How can you possibly discern a risk appetite with just asset and spending figures alone? The client might have health issues which mean they want to blow the vast majority of that retirement income early, and don’t need to maintain their current lifestyle for long.
Assume that the client wants to keep up their kids’ school fees sounds reasonable, but again personal circumstances can dictate times that wouldn’t be the case, and as their life progresses they may want to keep that mortgage or move house or move to an interest-only deal or do anything other than pay it off in the very specific period of 10 years that ChatGPT suggests.
FCA second test
Let’s move to the FCA’s second test: that the products are suitable based on those needs. The portfolio ChatGPT spits out is as follows:
Equities: 40%
Fixed income: 35%
Alternatives: 15%
Cash: 10%
Again, this arguably isn’t what the FCA is looking for. There isn’t really a link between the goals and why the investments help the client get there – bar that they fall within the made-up risk tolerance.
It also lists a few other recommendations: putting aside an extra £5,000 to ensure the client’s retirement income can meet their current lifestyle, considering placing some liquid assets into an education savings plan to secure their children’s’ education expenses, and increasing mortgage payments by £1,000 to reduce interest payments and ensure it’s paid off within a decade.
The same argument applies to these ideas: they are abstracted from what the clients actually wants to do, what makes sense to them outside of crude numerical calculations.
FCA third test
And what of the third FCA test: to document the potential risks? ChatGPT really has only one sentence to say about that:
“Based on our analysis of [name’s] situation, we believe the investment strategy and recommendations outlined in this report are suitable to achieve their long-term financial objectives with low to medium risk. However, we recommend that they review their financial plan on a regular basis and adjust it as needed to ensure that it remains in line with their evolving financial objectives.”
You can see how the regulator would think the client hasn’t been fully informed of the downsides after that.
This shows the crux of the problem of allowing such automation to dictate the contents of advisory reports. ChatGPT will be able to give you common reasons for recommendations. It might even be able to parse the web for the latest tax rates. But what it won’t be able to do is ask your client exactly what their aims and objectives are. What they want out of life. What they actually want their money to do for them.
Embedding AI in your processes
But that is no reason to abandon the tool altogether. On the contrary, it highlights where advisers can leverage ChatGPT’s strengths while being able to focus on their own.
What ChatGPT might be able to do is some formatting tasks. Imagine being able to take unstructured notes from a meeting, and turn them directly into the format, style and tone that best works for your firm, that is clearest to the client to and to the regulator.
Your paraplanner, in a few years’ time, might be the person best versed in that process. That would continue the worthy transition from glorified administrator to valued technician that that profession is currently experiencing.
The key here is the potential time saving as well as the clarity of formatting or semantics. So many of the man hours in an advice firm aren’t taken up by the meetings themselves. They are taken up by the documentation of those meetings, of communicating what comes next to the client and why.
So many advice firms already template reports to varying degrees. Using AI to generate some of those sections isn’t exactly a leap into the unknown.
Outside of suitability reports, the implications of ChatGPT for advice firms are huge. Some clients – though by no means all – like their weekly/monthly/annual updates from their adviser on how markets are doing. There’s no reason ChatGPT can’t just pull the best of the web together for you to illustrate the most significant changes that period, rather than business owners spending crucial time writing these newsletters and not concentrating on where the firm is going.
AI is no panacea for advisers. But don’t doubt it could have a whole bunch of uses for your firm.
Photo by Mojahid Mottakin on Unsplash
