Probably the most sophisticated AI/ML-based financial planning system in the UK is the Conquest Financial Planning proposition from Fidelity. This shows what should be normal in just a few years.
The adviser has the option to take complete control of the process or to be guided by the system, if you like a cruise control version, or if they wish Sam (the systems strategic advice manager) can take full control and effectively deliver a self-driven version of the advice.
Recordsure has developed a suite of AI technology to improve both the efficiency and the quality of call and meeting reviews. These can deliver huge advantages in enabling large numbers of file reviews at very low cost.
Clearly this has huge potential as firms aim to upscale reviews and checks in the run-up to consumer duty.
Talking about the success of the system group, Jo Smith, chief executive at Recordsure, points out that “our ML systems increasingly learn on the fly, from real reviewers simply doing their day job”.
In addition to pre/post-sale review use-cases, the technology can now be used to automatically triage in-or-out cases and calls in past business reviews. This can enable remediation exercises to be achieved in 20 per cent of the time taken by manual processes.
As we approach consumer duty, taking 80 per cent out of the time taken for file checks must be attractive.
Aveni are another organisation making major strides in this area particularly with the development of their “Human+” operating model designed to support sales effectiveness, meeting administration, compliance and ongoing service.
We are now seeing a significant growth in the number of automated and hybrid advice systems offered in the UK, these inevitably use AI. A recent study by FTRC to establish a benchmarking process for automated and hybrid identified seven such organisations and this is now being extended to include additional firms.
Our experience is that a growing number of advice firms are looking towards these systems as a key element of their processes to address consumer duty.
Risk of replacement
In its current iteration, Chat GPT is not the problem, it is a useful tool and could help us simplify many tasks, but it is limited. You do not have to push it too far to find its boundaries.
First, it does not know anything after the end of 2021, and while it can regurgitate knowledge it has no ability to cite sources for information. The real challenge is what comes over the horizon with versions four, five or six. Rumours suggest V4, may come as early as this quarter.