How much do you think the work of tax consultants will change using AI?
Fritz Esterer: This question is not easy to answer as the work of tax advisors varies from country to country. For example, tax advisors in France and Italy are more like tax lawyers who resolve specific tax problems and answer legal queries.
Their work will change significantly because they will be able to use AI and Legal Tech to carry out automated research activities for decision preparation and, thus, focus mainly on consulting. In the case-law-based legal area of Australia, for example, a tax consultant developed an AI system called Ailira, which prepares statements and decisions by using an extensive case database.
In Germany, the tax consultant prepares the tax returns for the clients and is the one who fulfils the reporting obligations. In addition, he also advises clients in tax law matters. On the other hand, the tax consultant in Brazil and all other Latin American countries is more of a tax accountant who mainly prepares tax returns.
The possible use of AI, therefore, strongly depends on whether one is more in a systematic or in a case-law-based legal area.
Specifically, the work of tax consultants will change in such a way that repetitive activities will be replaced by AI. The more repetitive work is replaced by the combination of technology, tools and AI, the greater the change in the composition of the tax functions and the work of the tax consultant.
In other words, the jobs are not eliminated, but the tax consultant can focus on other things and becomes more of the consultant?
Fritz Esterer: Definitely. This means, for example, that a German tax advisor will return to carrying out consultation services and be less involved in fulfilling declarations. In other countries, Legal Tech will certainly change the role of the tax advisor. I will even venture to predict that in countries where there are very comprehensive tax laws and where a case law system is used (such as the USA, Australia or Canada), the need for tax lawyers for simple preparatory activities will be reduced.
What expectations do you have regarding the use of AI?
Fritz Esterer: The value of AI and the two main benefits are that, on the one hand, more compliance security is gained. This can significantly reduce liability risks for the corporate management. This applies, in particular, to tax areas such as VAT, customs or transfer prices where mass data is processed, and even performs the most common duties which could potentially carry a high risk. On the other hand, AI makes it possible to prepare business decisions much better as more data can be processed in a way that humans could simply not replicate.
Another important aspect is the ability to connect all reporting obligations to the financial authorities. Every year, companies have to report dozens of topics to the tax authorities which may include items, in some cases, that are related to one another because they are generated from the same data. Some examples are the constructionwithholding tax, Intrastat reporting, declarations of permanent establishments or wage tax. Connecting the various reporting requirements can save companies a considerable amount of money. On top of this, it is also possible to make declarations even more consistent.
Could it even be possible that tax authorities are granted permanent access to data?
Fritz Esterer: That completely depends on how prepared various country tax authorities are willing to communicate with companies in a completely new way.
The desired aim would be for tax authorities and taxpayers to interact via an ongoing data exchange in real time, which in turn would give companies a certain degree of assurance.
Undoubtedly, this means we would have transparent taxpayers, but in my opinion, this will come with the digitisation process. Every company needs to be aware of this.
In which tax areas do you see the greatest short-term potential in the use of AI technologies?
Fritz Esterer: There is definitely potential where vast quantities of data are involved, such as customs, VAT and transfer pricing. Here, anomaly detection and process mining methods can be applied. AI can also be very effective with all types of tax deductions. On the one hand, I see less potential for income taxes, at least in the short term, because the amount of data available is too low. If this is the case, then it will only work with withholding taxes in selected cases.
In the short term, AI can help in terms of error analysis, compliance security and cost reduction. For example, in customs, it is particularly important to analyse deviations from certain rules stipulated in free trade agreements or in customs tariff databases. Regarding VAT, the main focus is on automatic checks of the monthly provisional returns. For these cases, appropriate AI solutions can be implemented very quickly in companies.
What are, currently, the biggest obstacles and practical limitations for the use of AI technologies in the tax area?
Fritz Esterer: On the one hand, there are objective topics such as the lack of data availability. On the other hand, there are also subjective aspects such as a lack of willingness to deal with AI that plays a role.
Since AI in taxation is still new territory for many companies and no extensive track records are available so far, there must also be a bit of pioneering spirit with the companies concerned.
Not to forget the cost aspect which is often a limiting factor. AI is not available "for free".
Companies have often made more progress with AI in other departments. One typical area of application is the supply chain, for example, where there has been much investment in AI. This also applies to the Accounting and Controlling department. However, only a few companies have used AI resources on the topic of taxation, so far. A corresponding necessity has not yet reached the awareness of the decision-makers, neither has the awareness of the fact that knowledge transfer in companies is necessary because what has already been developed for the supply chain can possibly even be used for the tax area.
This means, the tax function must not be viewed in isolation within a company to have the possibility of using AI?
Fritz Esterer: Tax functions must be increasingly integrated into general business processes. Once companies have achieved this step of consciousness, it will be much easier to deploy AI solutions in the tax area. For example, everyone should be aware that there is a closed-loop process in accounting for tax returns. The data is pulled from the Accounting department, the provisions are calculated and finally, the tax return is prepared.
What are the preconditions that need to be met to implement AI in the tax area?
Fritz Esterer: In order for companies to become AI ready, preparatory work is absolutely essential. This means that data must be prepared and structured, and processes must be standardized. Only then can the use of AI be successful. Above all, the tax functions themselves are required to create this precondition. However, this also requires resources and corresponding know-how.
Going forward, tax advisors need to think much more in terms of processes – but they are not generally trained to consider tax in this way. This is a major shortcoming and that is why there is a huge demand for tax advisors who are proficient in IT. As a result, there is high demand for relevant training opportunities and further study, or for training in AI, in general. A new type of job will, undoubtedly, be created – the digital tax advisor.
This is, therefore, the fourth core activity at WTS that we are also focussing on.
If we now think about the implementation, how do you think a successful introduction of AI could look like?
Fritz Esterer: As already mentioned, the first task is to capture all global control processes and then standardize them. Here, companies that have already introduced an internal control system (ICS) for taxation have a clear advantage. Within this framework, they have already defined control processes and controls as well as, if necessary, adapted the organisation. Basically, it is very complex to create the necessary data availability for AI solutions. Companies should know exactly what data they need for which AI application. Once this requirement is met, the data must finally be collected in a so-called tax data lake and structured using data analytics and business intelligence so that it can then be used for AI analysis, so that they can, subsequently, be interpreted into AI methods.
I think it is important to avoid making the mistake of just focussing on AI as it is an important component of digitisation. For the digitisation of tax functions, AI methods are used where it is appropriate.