Insurance Europe welcomes the opportunity to be involved in the ESAs consultation on the concept of automated advice and the current trends in the market, the potential benefits and risks involved, etc. This is the first step to ensure that any potential follow-up from the ESAs on this issue will only be taken if a need can be clearly demonstrated.
Furthermore, we acknowledge that the concept of automated financial advice and its prevalence varies greatly across the EU, and even between the different financial sectors, so it is an important starting point to have a common understanding of this concept based on existing legal definitions, in order to clarify the scope of the consultation and the discussion on automated advice going forward and to avoid confusing it with digitalisation of distribution of financial products in general.
With regard to the first characteristic, that the automated tool is used without (or with very limited) human intervention, we think that this needs to be qualified. As a principle, we believe that it is considered good practice to give the option to the consumer to have access to human advice should they want it at any point during the process. Even where the process can be completed entirely without human intervention, it is important that the consumer has the option during the process to contact an adviser if he wishes to do so. Indeed, some automated processes may even rely on some form of human interaction, for example in delivering the outcome of the process or querying the customer’s input. Greater recognition should therefore be given to the (positive) role human intervention can play in automated processes.
As regards the third identified characteristic, we are concerned with the broad application of the concept of advice to include anything that is perceived by consumers to be advice. We believe it is important to use legally robust definitions to avoid uncertainty and subjective interpretations of when advice is provided. The discussion of advice should therefore be aligned with the definitions of advice provided under financial services legislation. It seems inappropriate for the ESAs to disregard the existing ratified definitions of ‘advice’ in EU financial services legislation, but rather opt for a subjective and vague concept that lacks any real clarity over its potential scope.
A clear distinction is necessary between automated information/guidance and automated financial advice. There is an obvious difference between the provision of a personalised recommendation to a customer and other automated services that simply provide information in an interactive (and perhaps more engaging) way, allowing customers to make an independent choice. It is important when giving consideration to the concept of automated financial advice to only include those tools that involve the provision of a personal recommendation to a customer, and not those tools which serve only to provide information or guidance to customers, a point which seems to be at least partially recognised by the ESAs in paragraph 27.
It is difficult at this point in time to consider what possible characteristics automated financial advice tools may possess in the future. It is important therefore, given the relatively early stages of development of such tools, to not delimit the characteristics of such tools too strictly as this may inhibit further growth and innovation. It is possible that new characteristics may emerge over time, as technology is moving fast and reshapes the industry, and there should be sufficient room for this to take place.
We are aware that this is a growing area that is witnessing a lot of recent developments in the insurance sector, and receiving much attention from market participants across Europe.
The new distribution rules for insurance products introduced by the Insurance Distribution Directive (IDD) will also apply in the case of automated financial advice. The information conditions set out under Article 23 present an obstacle to the offering/development of automated advice tools and processes, as they require all advisory information to be provided on paper by default. Durable mediums other than paper or websites are only permitted under certain conditions (Article 23(2) IDD) and only “if the customer has been given the choice between information on paper and on a durable medium and has chosen the latter”. Providing the information by means of a website also depends on the customer’s prior consent. These rigid requirements are not suitable for the highly diversified business models in place and will lead to delays in the conclusion of contracts or even put their conclusion at risk. Moreover, it implies that information specifically designed for digital mediums is often poorly suited to be given on paper. One of the great advantages of digital communication is that there are much greater possibilities to target and layer the relevant information for consumers and in this way decrease the risk of information overload. Thus, any requirements for providers to give information designed for digital mediums also on paper represent an obstacle to the further development of automated advice and interaction with customers because it adds costs and even worse can result in the information given being less understandable and less accessible for customers.
It is therefore crucial that the regulation and rules on advice and information etc. are technologically neutral, meaning that the rules do not favour one medium (communication and/or interface) over another.
One of the key barriers facing companies that are considering offering or developing automated financial advice tools would be the upfront financial costs involved in creating and implementing such tools. It is important therefore that regulators and policymakers seek to ensure a stable and favourable regulatory framework, and conditions that will facilitate the further development of automated financial advice.
It is also important to ensure that guidance provided by regulators on the provision of advice is not simply copy-pasted across to the digital world from the face-to-face or the paper-based environment, without appropriately adapting it for the specific features of automated advice tools. Where such guidance is provided, it should be sufficiently principles-based so as not to impede companies from introducing innovative new solutions or create barriers to entering this market. Regulators could therefore play an important role in enhancing automation in financial advice by embracing a more omni-channel approach.
Digital exclusion should, for example, in some member states be tackled by stimulating investment in IT infrastructure and digital skills and education, thus unlocking a greater potential of automated tools. Enhancing digital literacy and the conditions in which it can flourish will contribute to further stimulation and consumer engagement with such tools.
Insurance Europe strongly believes that the development of automated advice will lead to significant benefits for consumers. We generally agree with the potential benefits to consumers that have been described in the text but would also wish to stress that there may be additional benefits in the future that cannot be perceived yet.
However, we would also point out that the use of automated tools, in general, may not be suitable for all customers, as some may not be comfortable with, or have easy access to, such automated tools, while others may simply prefer to receive human advice. We would therefore caution how easy all consumers would find using automated tools without any human interaction. For example, there may be instances when consumers need further explanation and so they should be given the opportunity to ask questions during the process.
We believe that automated advice could potentially help provide a more cost effective way for some consumers to access advice, and may be considered as an option by some who might otherwise not be in a position to afford advice. Digitalisation, generally, will reduce the search costs for consumers significantly. Some customers might also simply prefer online/digital services and interactions rather than the traditional human advisor interactions during regular office hours. However, at the same time, we also recognise that the use of automated tools may not be suitable for all customers, as set out in our response to question 6.
Automated advice tools may potentially afford consumers the possibility to take more time to process information, so as to better understand the advice that is being provided. Automated advice tools also help consumers to be more involved in the advice process, for example by assisting in uploading documents which are necessary for the advice to be provided.
In some cases, a consumer’s experience with automated advice will be the first step in the process, which in turn will increase the quality and depth of the following face-to-face or over the phone dialogue between provider and consumer.
Given that automation in financial advice is still in its early stages of development across the EU, we would note that there may also be other benefits which we are unable to foresee at this stage.
The key differences in terms of how the potential benefits for consumers will be distributed will largely depend on the automated financial advice model being considered and the nature of the decision that the tool is being used for. There are also differences because of the nature of the markets and the variances in financial products being distributed.
We generally agree with the potential benefits to financial institutions that have been described in the text but would also wish to stress that there may be additional benefits in the future that cannot be perceived yet.
Yes, automated advice models make it possible to introduce instant premium payment systems (such as via Apple Pay or Google Wallet etc) instead of collection of premiums via log processes with intermediaries.
We would assume that the differences in potential benefits are likely to be felt in markets / sectors where automation of financial advice is not currently commonplace. However, given that this is an evolving area (and this varies across EU member states and within financial services itself) we are not in a position to elaborate any further.
A number of the risks that are highlighted in the discussion paper are applicable to advice in general and are not specific to the area of automated advice. We would also argue that some of the risks presented should not even be considered as risks at all.
For example, R2 refers to the greater likelihood that consumers may incorrectly input wrong information into an automated tool. We would contest this assumption and would even suggest that the opposite is the case, as safeguards can be built into such tools to recognise inconsistencies or mistakes when inputting data and thereby prevent the consumer from continuing with the process.
Other ‘risks’ that are identified such as conflicts of interest and remuneration relate to all types of advice and not just to automated advice and these issues (risks) are already regulated under the respective pieces of financial services legislation (eg IDD, MiFID 2). Requirements under the IDD such as to act in accordance with the best interests of customers or to make suitable recommendations or present information in a form that is fair, clear and not misleading are applicable whenever advice is provided, regardless of whether it is automated or not.
Although one of the potential benefits identified for consumers is the easier access to a wider range of advice providers ‘including from other jurisdictions’, we would also point out that this could present a potential risk to consumers at the same time, as consumer knowledge will vary about differing regulatory and consumer protection regimes in other EU member states.
We agree in general with the description of the potential risks to financial institutions. However, for (re)insurance undertakings, many of the identified risks are common to the use of any IT system and are already dealt with under Solvency II requirements that provide for a risk management system.
We would also highlight the fact that R16 does not take into account relevant Solvency II provisions on outsourcing. The description of the risks related to liability allocation in paragraphs 83-85 fails to give any consideration to the strict regulatory framework for the outsourcing of functions under Article 49 of Solvency II, which stipulates that insurance and reinsurance undertakings remain fully responsible for discharging all of their obligations when they outsource functions or any insurance or reinsurance activities.
We believe that the use of automated advice tools will increase as the market evolves, particularly in the insurance sector. However, their use will not necessarily be uniform across all EU member states, so a sufficiently flexible approach should allow for this to be taken into account.
Given the benefits of the use of such tools, and the fact that the benefits greatly outweigh any risks involved, we would expect the development of automated advice tools to continue to grow in prominence. Further developments in artificial intelligence and cognitive computing may additionally offer great potential for application to the automation of financial advice.
In order to facilitate positive evolution in this area, further development of automated tools should be encouraged and care should be taken not to impede growth or stifle innovation through regulation that is obstructive or that becomes quickly out-of-date. It would not be advisable to establish prescriptive measures aimed at the standardisation of such tools and communications, particularly where companies are developing innovative, interactive and engaging ways to help consumers understand financial issues.
We believe that the potential benefits of automation in financial advice greatly outweigh any risks, for both industry and consumers, and so it is important to stimulate the development of tools that can further contribute to this.
As the use of automated advice in financial services varies greatly across EU member states, we believe it is important that a flexible approach is taken to any follow-up that the ESAs may consider. Some markets are significantly more advanced than others with regard to the development and implementation of automated tools, which should be taken into account in order to fully realise their potential and to avoid stifling innovation. This is a crucial point in order to ensure the further development of the EU Digital Single Market and in order to promote digitalisation as a driver for growth and increased competitiveness in the EU. Thus, it is very important that any initiative or follow-up in the area of automated advice (digitalisation of distribution) builds on this foundation and is undertaken with this same aim.
As already highlighted in our response to question 1, we regret that the ESAs have not considered using existing legal definitions of advice, which would give consistency and are more appropriate, and the fact that the paper adopts a focus on subjective consumer perceptions of advice. We believe it is important to use legally robust definitions to avoid uncertainty and subjective interpretations of when advice is provided.
The discussion of advice should therefore be aligned with the definitions of advice provided under existing financial services legislation, which has already undergone the appropriate legislative scrutiny. It seems therefore wholly inappropriate for the ESAs to ignore the explicit and precise definitions of advice that have been recently agreed under EU financial services legislation and to opt instead for a subjective and vague concept that lacks any real clarity over its potential scope. We would underline therefore that there is a clear difference between the provision of a personalised recommendation to a consumer and other services (or tools) that simply provide information in an interactive way, allowing consumers to make an independent choice.
While we understand the intention to explore the risks and opportunities associated with the automation of financial services, we believe that more prominence should be given to the fact that many risks are already mitigated by the existing framework of applicable EU and national legislation (eg IDD, PRIIPs, etc). For example, as far as the insurance sector is concerned, the automated financial advice tools considered in the discussion paper are, according to our understanding, within the scope of the newly entered into force IDD, which will provide for applicable rules whenever advice is provided, regardless of whether it is automated or not, and for all actors involved in a distribution activity regardless of the distribution channel used. Therefore, it would be advisable to await the full implementation of these requirements before assessing the need for further intervention in this field.