As to the assessment of the characteristics of automated financial advice tools presented in this Discussion Paper, we want to point out the following topics:
- The Discussion Paper restricts the current discussion on robo-advice to consumers: there are no provisions for non-consumer parties parties (like e.g. institutional clients).
- Although robo-advice may lead to a transaction, the advice in itself is not transactional. As the Paper explicitly mentions, partly automated processes have already been observed, but these processes would, under the current definition of robo-advice, fall outside the application of future legislative action.
- We would prefer to see a clear cross-sectorial definition of “advice”. For entities that are active in more than one sector, a uniform definition would prevent divergence of the meaning of “advice” across the sectors. We also propose the definition of “advice” to be rooted in existing legislation: on the one hand this would provide a basis for further discussion, on the other hand this would lessen the gap between current and future definitions of “advice”.
- A clear definition of “advice” would make clear whether “hybrid” tools are in scope of the discussion. It would also clear up the confusion which could arise when e.g. a consumer protection organisation develops a tool which would provide information that a consumer could perceive as “advice”.
- The level of specificity of the advice (para. 25) would not be a factor in determining whether the output is perceived as financial advice.
At the moment we do not see any further characteristics of automated financial advice tools.
We are aware of several automated financial advice tools, the most important ones are hinted at in the Discussion Paper.
At the moment we are considering offering automated financial advice tools, next to our existing hybrid financial advice tools (eg mortgage simulation combined with human intervention).
i) and ii) We are a long-established product provider in all three sectors.
The degree of human interaction in the tool we are considering is yet under discussion.
iii) We service both retail and non-retail clients.
v) We have not yet decided on who will develop the automated tool.
We see the following barriers:
- In a purely online environment the Know Your Customer procedures have to be re-assessed.
- In designing the algorithms, the miss-selling risk needs to be effectively mitigated.
- The consistency with existing channels within the entity or group of entities needs to be maintained.
- There is the question whether the customer is willing to pay for advice (in view of MIFID II).
- Due to the shift in distribution, the profit model will need to be revisited.
Benefits relating to costs:
- It may be the case that automating financial advice might make the advice more affordable. However, this assumption might come under pressure after MIFID II is in full force.
- We agree that through automated tools a wider range of consumers will have access to advice.
- Consumers will have be able to take a wider range of service providers to choose frominto consideration.
- Advice through automated tools may be provided faster, easier and in a non-consuming way. However, the accuracy and fitness of the advice may suffer if the underlying algorithm is not fine-tuned.
Benefits relating to quality of service:
- The advice through automated tools may be more consistent than advice provided through human interaction. Also, the consumers may always ask for an “upgrade” to full personalised advice. Given a wide enough range of products, consumers may be introduced to more fitting products than a human advisor may present.
- The algorithms, used to provide the advice, should have a real-time link to market data. In volatile markets the algorithms should still be able to give proper advice and adjust it if necessary.
- If automated tools would keep a record of the input the consumers provide, consumers might be able to compare results from other service providers and gain insights the products offered.
Advice through automated tools may support financial literacy and may lead to better financial inclusion (as observed in the lower-end of the market spectrum).
The degree of financial literacy may differ wildly across sectors. While most people will have some basic understanding of a current account, far fewer people will have that same financial literacy towards securization products.
Experiments with gamification increased the knowledge of the customers on the products concerned. Gamification is very educational.
In our opinion, a robo-advice tool is not a one-off investment. Investments will be needed on a continuous basis. There is a learning process of implementation to be considered (the customer journey may need to be adapted in view of future developments). Also, ever-changing regulation (e.g. in consumer protection) may make continuous investments necessary.
A specific regulatory framework for robo-advice may have the benefit that regulatory requirements will be consistently implemented across all three sectors. Given the current state of the regulatory framework, there is a risk of fragmentation in implementation.
As mentioned under Q8: a lot will depend on the customer journey: whether the customer is knowledgeable and self-directed or is in need of (a lot of) additional resources next to the automated process, will have an impact on the concept of robo-advice.
A more uniform approach (push for simplification) in the online process has drastically improved our client intake process.
- Because robo-tools may be used as instruments for financial education, the risk related to consumers having limited access to information may be easily mitigated, e.g. by including in the tool uniform terminology and providing ample explanation in the tool.
- Risks related to flaws in the functioning of the tool are not exclusively related to robo-advice. The process behind advice provided through human interpretation is equally unclear to a consumer.
- Some risks (e.g. consumers making unsuitable decisions because the tool does not prompt a trigger in the consumers to consider the impact of their decision) are easily mitigated through a correct set-up of the robo-tool. These risks are also not exclusively related to robo-advice: in a human advice-interaction the advisor may urge the consumer to make their decision quickly (pressure to meet targets).
- The herding risk described in the paper is also something not exclusive to robo-advice. Given the fact that there are no one-to-one solution for every consumer, consumers in very similar situations may indeed all decide to make transactions in the same product. The consistency of the content of the advice prevails, no matter the channel it is provided through.
- Fewer opportunities to access human advice is a market trend inherent to the current profit models of financial institutions and not a risk inherent to robo-tools.
Yes. Depending on the level of sophistication of the robo-tool (in connection to the financial literacy of the consumers), the potential risks might vary greatly. See Q8 and Q12.
Yes. Financial service providers need to make the needs of the consumers explicit and need to give concrete and appropriate advice. This is a complex process, which requires good knowledge of the customer.
When any of the described risks materialize, the customers may experience any of the adverse effects described.
Yes: pressure on profit models (see Q21).
Yes, very much. See Q8, Q12 and Q16.
Yes. There is increased competition (eg US Brokers) who are offering substantially lower prices in a non-sustainable business model (where the fees do not cover the costs). This puts a lot of pressure on fees structures.
Also, offering the same products via an alternative (robo-)tool may possibly lead to cannibalisation of existing channels, especially if the fee-threshold for robo-advice is lower than that for human advice.
Yes. Some reluctance from customers is expected (sensitive topics: consumers’ life-savings, personal data,…). There is however an important push from the “digital generation” towards more automated/online services.