Tax, Artificial Intelligence and the Taxexecutive Officer

Tax, Artificial Intelligence and the Taxexecutive Officer

AI: Artificial intelligence

TA: Tax Authority

TO: Taxecutive Officer 

A2B: Authorities-to-Business

Within 5-10 years time tax legislation will be converted into software codes. These can be run in a few seconds time to verify simple or complicated human tax advice. Beyond the obvious advantages of being quicker, cheaper, more accurate and coping with extremely bigger workload, AI has an uncompromising skill compared to humans, namely no inclination toward tax evasion or aggressive tax planning.

This might simply be interpreted and enforced via a new Asimov robotics rule, like "never evade tax or plan tax aggressively". For Tax Authorities a feasible feature of such AIs might be to install them on self-propelled/Turing-proof platforms and deploying at taxpayers' premises with or without human Tax Authority auditors. Those robots and their human counterparts might be both referred to as "Taxecutive Officers".

Of course the efficiency of robot TOs could be different to human ones, however their price (salary in case of humans) would surely reflect their qualities. At both ends of the two-way Authorities to Business communication channel staffing could be solved with either robots only, humans only or a mix of the two. Regarding the split between necessary competencies of advisory and compliance I would see a gradual shrinking of former, however ideally all TOs should perform both, at least on an acceptable level. Principally the shaping of (inter)national taxation systems and strategies has to remain in human hands, however consulting with human/robot TOs will certainly have an added value.

There is no doubt that the new role and impacts of TO will change our approach to the current ways of working with in-house tax departments, outsourced SSCs or 3rd party advisors. Starting with the trending topic of home working and its effects on productivity, it is easy to see that for a robot TO the definition of home office is irrelevant and due to the time consuming and expensive nature of commuting and maintaining a human office space, irrational too. Why would it then be different with regard to human TOs, who are treated as being at the same productivity level? Since future TOs will be much more efficient compared to present day tax professionals, the size of company tax departments will be decrease and employment patterns will benefit more from contractual relationships. Why could a TO not support more that one company if it has capacity for them, of course with confidential data treatment protection for all its clients?

What will the optimal time split of a TO be? Currently 70-90% of a TO’s time is spent with operation/maintenance/audit support of averagely efficient reporting systems and the rest can be used for managing advisory/miscellaneous ongoing issues. In an ideal world the split should look like 1/3 reporting systems operation, 1/3 advisory/miscellaneous ongoing and 1/3 self-improvement/optimisation/tax planning. Big4 firms and TAs will likely move to this new era in the first lines, however new autonomous TO networks might also grow up securing the qualified and affordable taxation supporting capacities for companies. For moderate prices TAs could offer the services of their own TO networks, which could be a great deal for certain scope of SMEs, enabling them to focus on their main revenue sources. Due to the common service standards, TO assistance interchangeability will be smooth among various quality levels and scenario complexities, meaning eg. that a company is usually purchasing a robot TO from TA, while every third month a human TO comes for two weeks and once a year a high prestige big4 human/robot TO team doublechecks the taxation systems, etc.

A changing workforce needs changing tools too. In recent years hype around semi-automatic VAT compliance softwares has been booming and tax technology positions are gaining momentum. This brings different approaches and formats to the market at the very beginning of a longer route which should be a unified one "unfortunately". Accordingly those service providers maybe celebrate too early about their "robust", "unmatched" and "high-tech performances, since we are at the beginning of an evolution and industry leaders should think 2-3 steps in advance already. Thus a comprehensive end-to-end taxation software has to have a central, self-checking, big data analytics module, which can manage:

(a) one-click compliance: With respective knowledge of legal rules of current/previous taxation years in necessary jurisdictions, where submission and financial settlement actions would be both triggered with that one click. Prior to that calculated tax positions could be comprehensively but quickly checked/amended within the central analytics module. Details of amendments automatically recorded in dedicated databases, controlled by analytics module.

(b) masterdata quality enhancer tools: Since the Garbage In Garbage Out principle will not lose its significance, the quality of tax related accounting masterdata can boost efficiency of tax procedures. Instead of long waiting in-house or outsourced remedy for missing/incorrect parameters it is more efficient to continuously monitor with simple, non-stop search algorithms all accounting entries/databases and reacting to their findings immediately. Doing so, burden can be taken off compliance procedures.

(c) bulk correction of accounting entries and accounting entry wizards: In case of complicated issues adjustments should be performed in transparent and structured way even re several thousands of records the same time. Number of errors might be reduced right at the posting with accounting entry wizards, which would pre-filter entries and even throwing back if having poor data quality.

(d) direct data extraction towards TA: Technical prerequisites of taxation big data handling on TA side will be soon fulfilled and as a consequence tax will be calculated by them for every taxpayers instead of slow and unreliable self-declarations. The trend of asking more and more data from companies points in that direction also.

(e) tax audit connection hotspot: Serves as a unified docking station for the usual TA audits enabling quick data exchange between the parties. A practical additional feature might be launching pre-audit stress tests demonstrating integrity of company tax systems and their readiness for an audit. Trend-analysing heat maps make it possible to identify unexpected volume changes, implying potential system malfunctions.

(f) assistance in tax planning through data visualisation: In order to maximise benefits of applicable taxation options, visualising revenue/cost streams, cash flows and corresponding tax charges per countries of business presence can massively contribute.  

(g) access rights to company taxation data: Primary function of this is the quick but controlled change between different TOs, plus acting as the first line of defence against cyber threats.

On a sidenote, the possibilities of process automation and optimisation tools (beyond taxation, developing accounting/closure/controlling/forecast/etc assisting automatisms) are almost unlimited and for sure there are such, which could not even be thought about yet as the technological development is not there. We might be in a similar situation to the frog of Frithjof Schuon who denies the existence of mountains as they are illogical, according to the puddle knowledge base. I like that dilemma, since it is not implying that the frog is stupid, simply the current circumstances (ie the puddle) are strictly limiting its creativity.

Regarding the taxation big data streaming into TA databases it would be logical to run non-stop Google-like search algorithms looking for tax shortages and potential subjects of tax audits. Being on the safe side, performance of first layer algorithms should be supervised by a second layer as well. As mentioned already the topic of different formats it is likely that in the first few years a kind of format war might be escalated (like VHS vs Betamax or Blu-ray vs HD DVD), however the final say will come from TAs. In order to avoid massive format conversion tasks, TAs should be aware of latest available developments and treat the selection procedure with high priority. This TA contest might be easily won with the quickest reaction speed, eg in the post-Brexit era the UK TA does not have to reach political/technological consensus with any other country and thus could build the most effective, low tax rate, technology based and user friendly taxation system. It also should not be surprising if the tax charge will become payable in forms of electricity or computing capacities. 

In conclusion, technology will bring about equality and let taxpayers concentrate on what really matters: core revenue activities. Simultaneously, the tax profession will get back to basics to restructure itself in a technology based, unified, supportive and efficient format. One thing to remember: we should not grow business in order to pay taxes but pay taxes in order to have the chance of growing our business.

This article was written by Balint Batonyi. He has 15 years' tax experience with a focus on Eastern Europe.