Yes, more chat about AI. But given the questions we’re hearing from financial advisers, we really think it’s necessary! More and more advice firms are adopting AI. If you know us at all, you know we’re mostly friends with the robots – check out our AI hot topic for the proof – but it’s really important to understand appropriate use. So let’s answer the burning question. Can you use ChatGPT (or Copilot, or Claude…) for client work?
Let’s start with what’s possible. AI is helping loads of businesses across the spectrum of industries. It’s broadly helpful for:
And adviser-specific AI tools are also helping firms:
(Watch back webinars from our AI Spotlight series here to see how some of the tools work.)
We need to be realistic about compliance risks around AI; they do exist.
For firms using AI to process data and complete tasks, security and accuracy take centre stage.
From a security perspective, it’s paramount to protect against GDPR breaches, confidentiality loss and unauthorised use by staff. Remember that even partial information can identify a client.
Accuracy-wise, consider a lack of audit trails and hallucinations – that’s false information from a tool whose generative predictions have fallen foul of reality. In other words, AI can sometimes make something up if there’s a knowledge gap.
The key to getting it right (read: compliant) is to prioritise AI tools that are both secure and accurate, proportionate to your use.
Generally safer uses include using AI to draft generic communications, create meeting agendas, brainstorm ideas and improve your writing (don’t chuck it all through ChatGPT without checking though, we see you with your em dashes). We still highly recommend using a paid license and choosing watertight security settings.
Higher risk uses include uploading suitability reports, sharing client fact finds, inputting personal data or asking for client-specific recommendations.
Using general AI tools for these higher risk activities is not recommended.
Instead, choose an industry-specific AI tool that’s designed to do the job you want to give it. You should also look for ensured data confidentiality, systemic robustness and fairness and per the FCA, clear human accountability at all times.
If you choose to use AI, we highly recommend taking the time to take stock, and prepare properly:
The answer, we’re afraid, is no. But you can use other AI tools, and we can show you how to choose (and use) safely. Our AI hot topic area is a good place to get started with articles, webinars and how-tos.
With so many powerful tools out there, it’s an exciting time for productivity – but hugely important that firms apply governance before adoption.
Would your firm’s AI usage stand up to regulatory scrutiny?
If you’re in any doubt about how to integrate AI or continue using it safely, book a chat with our compliance team today – we’d love to help.
Alanis is a compliance consultant at Verve with an AF7 qualification and deep technical expertise across all areas of quality assurance and regulation.
.png)
.png)
.png)
