New guidelines on robo-advice and (semi-)automated asset management

What is this about?

Professional market participants are developing automated advice services for various financial products. The usage of so-called “robo-advice”, in which a self-guided digital service provides (partly or fully) automated advice with minimal human interaction, is increasing. Moreover, there is a growing practice of (semi-)automated asset management in the investment community. The current Dutch legislation and regulations do not make a distinction between (semi-)automated and physical services. Therefore, both services have to meet the same standards and are subject to the same conduct of business rules. However, there are specific areas with regard to (semi-)automated services which need consideration.

In the Guidelines specific key focus areas are outlined and guidance is provided; in particular regarding the quality of the advice and the duty of care. The first Guideline (Leidraad invulling zorgplicht (semi-)automatisch vermogensbeheer) relates to (semi-)automated asset management services provided by (amongst others) investment firms. In addition, the second Guideline (V_isie op roboadvies: kansen, zorgplicht en aandachtspunten_) provides guidance and examples of good practices for parties offering or advising on complex financial products such as mortgages and invalidity insurance by way of robo-advice. This publication will focus on the first Guideline.

Which parties should care?

Investments firms, banks and managers of investment funds, that (are considering to) provide any form of (semi-)automated financial services to retail investors ought to be aware of the Guidelines. Furthermore, the Guidelines are of importance for (non-regulated) suppliers of technology which is used for (semi-)automated financial services.

What should they know?

Although the Guidelines are not binding, the AFM expects the abovementioned parties to examine the contents of the Guidelines and further develop their practices and procedures based on the recommendations included in the Guidelines. For the Guideline on the fulfilment of the duty of care for (semi-)automated asset management services these are:

Operational measures

• adequately organise their business operations by developing algorithms and online tools in which the outcome primarily serves the interest of the client
• only use algorithms and online tools if they are capable of effectively supervising, monitoring, testing and updating such tools
• maintain sufficient knowledge and expertise inhouse when outsourcing certain activities related to the provision of (semi-)automated services, and
• install an independent and effective compliance function.

Information from clients

• use a “smart” questionnaire for a client survey, including personalised exit-, control-, and challenge questions
• have a detailed policy in respect of keeping client information up-to-date
• record client information with care (audit trail)
• use objective external sources to obtain more reliable information on clients
• consider quantitative methods for the survey of risk acceptance
• consider using data on clients and insights on behaviour in the interests of clients, and
• ensure client data is protected and processed correctly.

Information to clients

• pay extra attention on clear information in order to make it as comprehensible as possible for clients
• inform clients regarding the investment services provided and offer assistance in choosing a specific type of investment service
• apply information obtained on customer behaviour, client data and persuasive techniques in the interest of the client
• use the possibilities of digital communication in the best way possible, and
• inform their clients on an ongoing basis before and during the provision of the service.

Product governance

• assess the target group for the provision of the relevant (semi-)automated service
• ensure that the concept of (semi-)automated asset management aligns with the target group of the products which need to be distributed
• determine and evaluate the target group and the strategy on distribution on a regular basis in addition to the individual test on suitability
• apply self-learning algorithms restrictively, and
• involve various specialties in the design and evaluation of the (semi-)automated service provided.

Any further thoughts?

The AFM has reiterated that - in principle - (semi-) automated services are subject to the same rules as physically provided services (eg in a face-to-face setting). The Guidelines serve as additional illustration of how to apply certain legal requirements and principles in an online or (semi-)automated setting.