Poorly designed robo-advice could lead to “systemic mis-selling” the head of strategy and competition at the Financial Conduct Authority has warned.
Bob Ferguson said an automated advice model could help mitigate some of the risks associated with human advisers and managing a large salesforce but ultimately it would depend on the model.
He said: “The design of the model is crucial – a poorly designed model could lead to systemic mis-selling.”
Mr Ferguson said managing risk was ultimately the responsibility of the individual firm and its senior management.
He said the FCA’s rules are “technology neutral” and do not change regardless of whether a company has outsourced a function to a technology provider.
In a speech in London Mr Ferguson said: “Managing this risk, and others, is ultimately the responsibility of the firm and its senior management.
“This responsibility isn’t reduced if the firm uses third party suppliers to help with the technology part – it rests with the firm offering the system.
“And if, for example, your method of profiling the risk appetite of the client is flawed, it is no more defensible because you sourced the model externally.”
Mr Ferguson listed some of the themes the FCA has encountered in its work on robo-advice.
One was the expectation that a typical model might emerge which is a hybrid of human advice with automated technology, instead of pure robo.
He said: “We’re not ready to answer with any certainty whether this will be the case as we look forward.
“But it’s true that one of the areas of feedback that the Advice Unit has been asked about has been combining human intervention with automated advice processes.”
One of the issues advisers faced, Mr Ferguson added, was that people tend to search for products such as Isas or pensions rather than advice or guidance.
He recommended that advisers need to be clear about the nature of the service they offer, in terms of whether it is advised or non-advised and how wide its scope is.
The FCA is currently carrying out an assessment of the suitability of automated advice, with the first phase expected to be completed in the first quarter of next year.