Thanks to IT and dig­i­tal­i­sa­tion, automa­tion has final­ly made it to finance and tax departments.

Dig­i­tal­i­sa­tion has been mak­ing an increas­ing num­ber of pro­fes­sions more effi­cient than ever. Not only are they to be auto­mat­ed in the future, but this process is already under­way: faster, eco­nom­i­cal in terms of staff, more pre­cise…. What is going on in the tax con­sult­ing and audit­ing sec­tors? How are the many process­es at work being man­aged and direct­ed by means of Big Data, dig­i­tal sig­na­tures and AI? What pro­fes­sion­al pro­file will this sec­tor be requir­ing? Is unem­ploy­ment a threat or will a new­found free­dom allow for oth­er goals and tasks to final­ly be tack­led? What might those new aims be?

In her Duet inter­view with Chris­t­ian Klein, tax and audit­ing con­sul­tant, Dr Cal­daro­la, author of the recent book Big Data and Law, chats on recent devel­op­ments in the field of tax con­sult­ing and auditing.

For years peo­ple have been able to go to a book­store and get soft­ware allow­ing them to do their own income tax returns. More recent­ly, tax con­sul­tants no longer give in paper ver­sions of income tax returns. Instead, elec­tron­ic ver­sions are being sub­mit­ted. What changes have you already noticed regard­ing dig­i­tal­i­sa­tion and automa­tion, what is cur­rent­ly being worked on and what will hap­pen in the future? 

Chris­t­ian Klein: Dig­i­tal­i­sa­tion is a trend which has been devel­op­ing for a long time, but its progress has been accel­er­at­ing in the last few years, espe­cial­ly with­in finance depart­ments. On the one hand, more and more data are being col­lect­ed and stored. The vol­ume of data held with­in com­pa­nies is grow­ing in leaps and bounds and dou­bles every two years. Part of the rea­son for this devel­op­ment is that busi­ness mod­els are them­selves becom­ing more and more dig­i­talised, and mass amounts of data are being pro­duced. Owing to this change, com­pa­nies are more will­ing than ever to trans­form their inter­nal process­es. At the same time, how­ev­er, only a small amount of this data is actu­al­ly being used. The con­sult­ing indus­try has react­ed to this sit­u­a­tion by devel­op­ing solu­tions which enable com­pa­nies to man­age this data more effi­cient­ly and use it for their busi­ness pur­pos­es. AI has still not been accord­ed the impor­tance it deserves. This will sure­ly change in the long term, par­tic­u­lar­ly when it comes to recog­nis­ing pat­terns and pre­dic­tive analy­ses. At the moment, how­ev­er, dig­i­tal­i­sa­tion projects are still main­ly focused on basic mat­ters such as process automa­ti­sa­tion, data har­mon­i­sa­tion and the improve­ment of data qual­i­ty which already direct­ly gen­er­ate val­ue with­out the aid of AI.

Cur­rent­ly, we can observe that part of our job involves accel­er­at­ing and qual­i­ta­tive­ly improv­ing how day-to-day busi­ness oper­ates and how rou­tine tasks are car­ried out. These include prepar­ing finan­cial state­ments and income tax returns, cre­at­ing an inte­grat­ed inter­nal and exter­nal KPI and busi­ness per­for­mance report­ing as well as ful­fill­ing com­pli­ance require­ments and mon­i­tor­ing busi­ness process­es in terms of what is finan­cial­ly relevant.

A core aspect of tax­a­tion is min­imis­ing risks. Laws con­cern­ing sales tax­es, in par­tic­u­lar, have a lot of poten­tial to be dig­i­talised, as the rel­e­vant cri­te­ria used to eval­u­ate and col­lect sales tax­es, which are of course sub­ject to the var­i­ous nation­al laws, can be eas­i­ly imple­ment­ed by a so-called tax engine. This is an excit­ing devel­op­ment, par­tic­u­lar­ly for com­pa­nies in glob­al trade. In this way, even with the tech­nol­o­gy we already have at our dis­pos­al, we are able to have sales tax­es auto­mat­i­cal­ly ver­i­fied at every invoice receipt regard­less of where the trans­ac­tion has tak­en place or in accor­dance with which legal sys­tem. This type of automa­tion brings risk cov­er­age with regard to declar­ing sales tax­es to a new lev­el, when a sub­se­quent and only sam­ple-based invoice ver­i­fi­ca­tion is replaced by a ful­ly auto­mat­ed invoice-relat­ed ver­i­fi­ca­tion process in real time. A tax engine is, there­fore, a pow­er­ful tool to low­er risks, reduce monot­o­nous tasks and save costs.

The dis­clo­sure require­ment accord­ing to the Euro­pean Sin­gle Elec­tron­ic For­mat (ESEF) is one exam­ple that illus­trates new oblig­a­tions regard­ing exter­nal finan­cial reports. This for­mat is a stan­dard­ised elec­tron­ic report valid through­out Europe which is meant for cap­i­tal-ori­ent­ed firms, whose annu­al finan­cial state­ments must be report­ed via a defined tax­on­o­my and are thus machine-read­able. Even the audi­tor must issue a dig­i­tal audit opin­ion on this new report­ing. Incor­rect ESEF report­ing can even lead to a qual­i­fied audit opin­ion or a refusal of the audit report. This reg­u­la­tion means that the EU Com­mis­sion has cre­at­ed a new sit­u­a­tion which will sure­ly lead to clos­er coop­er­a­tion between com­pa­nies and reg­u­lat­ing authorities.

It is only a mat­ter of time before the tax author­i­ties increase their data require­ments. They already have com­pre­hen­sive access rights to finan­cial­ly rel­e­vant data, a right which the author­i­ties have increas­ing­ly employed when it comes to tax audits. In such a case it is impor­tant to be able to have the data avail­able as effi­cient­ly as pos­si­ble and to be in a posi­tion to explain the auto­mat­ed process­es in a com­pre­hen­si­ble man­ner. The use of new IT solu­tions for the analy­sis of mass data and the ver­i­fi­ca­tion of tax returns still has a long way to go, but it is already clear that such a devel­op­ment will sure­ly trans­form the way we com­mu­ni­cate with the tax author­i­ties. In oth­er coun­tries, for exam­ple, com­mu­ni­cat­ing with the tax author­i­ty is auto­mat­ic, mean­ing that a busi­ness trans­ac­tion between two com­pa­nies trig­gers a data report via a stan­dard­ised inter­face to the tax author­i­ty at the same time. Obvi­ous­ly, those coun­tries do not have sub­se­quent annu­al income tax returns as we know it.

We can observe this trend in oth­er areas, such as audit­ing firms expand­ing their ambi­tion to be a long-term spar­ring part­ner. It is pos­si­ble that the annu­al audit will devel­op into a pro-active and asso­cia­tive process by cre­at­ing and main­tain­ing an inter­face between the Enter­prise Resource Plan­ning (ERP) of the rel­e­vant com­pa­ny and the audit­ing soft­ware of the audi­tor. This offers total­ly new pos­si­bil­i­ties in the sense of per­ma­nent cov­er­age real-time audits. These are not, in fact, new ideas. Rather, car­ry­ing them out has not been pos­si­ble owing to the lack of stan­dard inter­faces, soft­ware solu­tions and the desire for change in com­pa­nies, which are all nec­es­sary prerequisites.

To sum­marise, we can say that our pro­fes­sion is at the moment sim­ply not pre­pared for these devel­op­ments, just as the ERPs and IT solu­tions being used today are also not up to the task. Medi­um-sized com­pa­nies, in par­tic­u­lar, are fac­ing some chal­leng­ing times, but also promis­ing ones. The real dif­fi­cul­ty lies in recog­nis­ing what is now pos­si­ble thanks to dig­i­tal­i­sa­tion and then suc­cess­ful­ly apply­ing these new devel­op­ments with­in the company.

If more and more is being auto­mat­ed with­in your field, will the user rely on the result obtained from machines? Does this mean that pri­or knowl­edge and abil­i­ties will dis­ap­pear, like doing math in your head has been replaced by cal­cu­la­tors or read­ing a map by nav­i­ga­tion devices? Or will new abil­i­ties devel­op in their place? If so, which ones?

Dig­i­tal­i­sa­tion is meant to pro­vide us with an improve­ment in qual­i­ty and effi­cien­cy as far as our dai­ly work is con­cerned. Gen­er­al­ly speak­ing, with regard to finance and tax depart­ments, the most time-con­sum­ing tasks are usu­al­ly rou­tine ones, like col­lect­ing data and then prepar­ing, check­ing and report­ing them. On aver­age, account­ing offices spend 70% of their time col­lect­ing and prepar­ing data and only 30% remains for the exe­cu­tion of the orig­i­nal task. In this regard, dig­i­tal­i­sa­tion can free up valu­able time.

If we con­tin­ue using the exam­ple of tax engines, those who use it no longer have to search for the prover­bial nee­dle in a haystack by ran­dom sam­pling dur­ing a month­ly tax val­i­da­tion. Anom­alous cas­es will be auto­mat­i­cal­ly passed on to tax experts via a work­flow, who can then exam­ine them in detail. Those tech­nolo­gies can help to real­ly improve day-to-day busi­ness in these areas and can pro­vide more time for finan­cial experts to work on things that require more effort. Data will be avail­able in real time and are to be used for tax pur­pos­es in order to iden­ti­fy risks at an ear­ly stage and to make data-based deci­sions.  It is not that tax experts will be replaced, but rather they can spend their time focussing on val­ue-cre­at­ing activ­i­ties includ­ing an expan­sion of these activ­i­ties. Using new dig­i­tal tech­nolo­gies changes the duties and func­tions of the tax department.

Will this mon­i­tor­ing result in a new type of pro­fes­sion just like soft­ware pro­gram­ming for automa­tion in finance? Will it even be pos­si­ble for a group of peo­ple to under­stand the many steps and cor­re­la­tions, all of which is done by AI in a moment’s work?

That is a very impor­tant aspect which is often not con­sid­ered enough. The less man­u­al activ­i­ty exists in the process being worked on, the more tech­ni­cal reg­u­la­tions come into play. These reg­u­la­tions are, for instance, deci­sion rou­tines which have been pro­grammed and which auto­mat­i­cal­ly run in the back­ground. The more we rely on these auto­mat­ed process­es, the more impor­tant it is for us to under­stand them at a con­cep­tu­al lev­el and to check them at the tech­ni­cal. Oth­er­wise, there is the risk of high­ly auto­mat­ed process­es being used which can be seen as a mate­r­i­al risk to the true and fair view of the finan­cial posi­tion and per­for­mance of a com­pa­ny due to the mass trans­ac­tions tak­ing place. These reg­u­la­tions must, there­fore, be pro­fes­sion­al­ly exam­ined and approved when they are still being devel­oped. Fur­ther­more, con­trol mech­a­nisms need to be imple­ment­ed and reg­u­lar­ly mon­i­tored with­in the rel­e­vant process­es. Inter­nal con­trol sys­tems are fun­da­men­tal to adher­ing to com­pli­ance require­ments. This is an area to which tax con­sult­ing and audit­ing firms strong­ly ori­ent them­selves. Audit­ing firms have spe­cif­ic experts for IT-rel­e­vant finan­cial process­es work­ing on their audit­ing teams. The inter­re­la­tion­ship can thus be seen to take on even more impor­tance, togeth­er with the sig­nif­i­cance of IT expertise.

In my opin­ion, the term AI is cur­rent­ly still very much dri­ven by a mar­ket­ing per­spec­tive. Data are being col­lect­ed, struc­tured and processed. These data are then sub­ject to prob­a­bil­i­ty cal­cu­la­tions in a fur­ther step in order to come to cer­tain pre­dic­tions. In short, we are talk­ing about prob­a­bil­i­ty and sto­chas­tics. A com­plex busi­ness, to be sure, but a very struc­tured one which can be con­sid­ered ratio­nal­ly, doc­u­ment­ed and, of course, verified.

In today’s world, many com­pa­nies are not real­ly con­cerned about the prob­a­bil­i­ty ele­ment. While there are still areas which use solu­tions capa­ble of mak­ing pre­dic­tions, such as the auto­mat­ed plan­ning of what is need­ed in the con­sumer goods and trade indus­tries, most com­pa­nies, how­ev­er, empha­size oth­er aspects. These include, in a pre­lim­i­nary step, work­ing with mass data, struc­tur­ing them, pro­cess­ing and report­ing them in a mean­ing­ful way. Today, this is where the approach to con­sult­ing begins, an area which is already gen­er­at­ing a lot of value.

With increas­ing automa­tion respon­si­bil­i­ty dis­ap­pears because peo­ple rely more and more on machines and are no longer able to have a com­plete overview of the process in ques­tion because it has become too com­plex.  We can com­pare this with pro­duc­tion lines which first appeared with the advent of the Indus­tri­al Rev­o­lu­tion. Even in this sit­u­a­tion it is no longer pos­si­ble for a sin­gle employ­ee work­ing at the assem­bly line to know where the raw mate­ri­als and assem­bly parts have been pur­chased, how they have been processed or how and where they have been sold. In this way, indi­vid­ual work­ers no longer feel any sense of respon­si­bil­i­ty for the end prod­uct.  Respon­si­bil­i­ty, how­ev­er, is quite sig­nif­i­cant in your pro­fes­sion. Will the con­cept of respon­si­bil­i­ty change?

The use of new tech­nolo­gies is of course accom­pa­nied by big­ger chal­lenges and more com­plex sit­u­a­tions. The sta­bil­i­ty of the sys­tem and data secu­ri­ty are also giv­en greater sig­nif­i­cance. The chal­lenge lies in deriv­ing what is known about legal require­ments in tax and account­ing and to suc­cess­ful­ly imple­ment and mon­i­tor them using new tech­nolo­gies in com­pa­ny-spe­cif­ic IT solu­tions. Respon­si­bil­i­ty for data and report­ing process­es is increas­ing because they are the basis of com­pa­ny tax­a­tion and finan­cial reports. More is being required of employ­ees, their com­pe­ten­cies and respon­si­bil­i­ties, but also on the organ­i­sa­tion­al struc­ture and super­vi­so­ry bodies.

For exam­ple, the IT domain bears a com­pre­hen­sive respon­si­bil­i­ty for secur­ing the nec­es­sary infra­struc­ture, cre­at­ing the required data trans­mis­sion as well as for adher­ing to IT gov­er­nance. With regard to the skills need­ed for analy­sis and report­ing, how­ev­er, a robust devel­op­ment of exper­tise in finance organ­i­sa­tions can be observed. Apart from under­stand­ing the process, a cer­tain basic com­pre­hen­sion of automa­tion is need­ed as well.  Nev­er­the­less, this does not mean that a tax expert must now also devel­op data banks. On the oth­er hand, what I have noticed to be valu­able and what I myself apply is estab­lish­ing “com­pe­tence teams” with­in the finance organ­i­sa­tion. Apart from process man­agers and tax experts, there are also BI experts and ana­lysts, who them­selves have a basic knowl­edge of book­keep­ing and under­stand the issues involved in the fields in question.

But even the way we work togeth­er requires changes in the struc­ture of the organ­i­sa­tion. A strict­ly ver­ti­cal hier­ar­chy quick­ly reach­es its lim­its. What has often been found to be suit­able is a BI com­pe­tence cen­tre which then coor­di­nates coop­er­a­tion between IT and finance with regard to the top­ics and resources in ques­tion. A com­pe­tence cen­tre is the right place for process­es to be man­aged end-to-end by experts from rel­e­vant depart­ments, for a high lev­el of respon­si­bil­i­ty for gov­er­nance, for estab­lish­ing work­flows and tool solu­tions and for shar­ing cross-depart­men­tal knowledge.

Respon­si­bil­i­ty is increased fur­ther via the devel­op­ment of con­trol mech­a­nisms which are inte­grat­ed with­in the process­es them­selves. The inter­nal con­trol sys­tems must be devel­oped in con­junc­tion with automa­tion and are to be con­trolled by means of inter­nal revi­sion as well as an audi­tor. It is clear that dig­i­tal­i­sa­tion will trans­form the work and indeed the pro­fes­sion itself of revision.

A lot of work is done on paper in your pro­fes­sion, whether we are talk­ing about invoic­es, bills of lad­ing, con­tracts and much more. A data pro­tec­tion agent might think that this is all “unstruc­tured data“, for the var­i­ous data can­not be found in spe­cif­ic fields, but rather on a sin­gle or many pieces of paper. Fur­ther­more, not every­one writes their invoic­es, bills of lad­ing, con­tracts etc. in the same way, mean­ing that indi­vid­ual data are to be found in var­i­ous places and in var­i­ous for­mats. How­ev­er, a basic require­ment of dig­i­tal­i­sa­tion is for the data to be struc­tured. How is this lat­est chal­lenge being managed?

The het­eroge­nous nature of data qual­i­ty and data source are indeed one of the main chal­lenges in imple­ment­ing any strat­e­gy involv­ing dig­i­tal­i­sa­tion. This can occur, to name one exam­ple, if the sys­tems being used with­in a com­pa­ny do not “com­mu­ni­cate” with one anoth­er, a sit­u­a­tion which can lead to sys­tem breaks. Inter­faces are not always avail­able, and many com­pa­nies already have more than one ERP sys­tem at work because they have been allowed to devel­op in this way over the course of time. In the end, it all depends on the pre-exist­ing con­di­tions, and not every­thing has to be changed and devel­oped from scratch on a green­field. A tried-and-true method has been to see which infor­ma­tion is already avail­able by means of the com­pa­ny using a data ware­house solu­tion and then expand­ing on it for tax require­ments. At this point, things start to get tech­ni­cal. Depend­ing on what strat­e­gy has been cho­sen, one might form a “tax data lake”, which is based on var­i­ous data sources. These sources might be sev­er­al ERP sys­tems, data banks, but also Excel report­ing pack­ages. These data are then pre­pared for tax pur­posed using so-called ETL lines (Extract, Trans­form, Load) and made visu­al­ly avail­able via some sort of report­ing solu­tion. In this way, var­i­ous quite dis­tinct data struc­tures can often be har­monised. These solu­tions do not depend on any one sys­tem, offer com­plete­ly up-to-date report­ing sim­ply by press­ing a but­ton and, depend­ing on the type, can even have the option of direct analy­sis. These are all Busi­ness Intel­li­gence solu­tions which rely on a close col­lab­o­ra­tion with the exist­ing IT.

Apart from how data are set up and processed, a basic prin­ci­ple is also required which spec­i­fies how mas­ter data are to be man­aged. Many com­pa­nies are not aware of the ben­e­fits of con­trol keys and Incoterms, even though the larg­er ERPs, in par­tic­u­lar, inher­ent­ly have this func­tion­al­i­ty already.

Dig­i­tal­i­sa­tion usu­al­ly leads to stan­dard­i­s­a­tion but also to less cor­rup­tion. Owing to the increase in com­put­er-sup­port­ed doc­u­men­ta­tion, how will indi­vid­ual cas­es be han­dled which do not fit the norm?

By means of dig­i­tal­i­sa­tion, for exam­ple, an algo­rithm can be used to cal­cu­late an inde­pen­dent and expect­ed val­ue and includ­ed into a pro­cure­ment event in the ERP sys­tem. When the asso­ci­at­ed incom­ing A/P invoice dif­fers sig­nif­i­cant­ly from this val­ue, an account­ing clerk will auto­mat­i­cal­ly receive a mes­sage. What this shows is that case pro­cess­ing is car­ried out using a pre­de­fined work process, so that there is lit­tle room for manipulation.

In terms of sales and mar­ket­ing, BI analy­ses can aid in close­ly mon­i­tor­ing rev­enues which were gen­er­at­ed and report­ed by exter­nal agen­cies as well in test­ing them for their plau­si­bil­i­ty. This test is done, to name an exam­ple, by com­par­ing the data report­ed by an exter­nal source with a val­ue inter­nal­ly cal­cu­lat­ed using so-called “shad­ow billing”.

Even tax author­i­ties have begun coop­er­at­ing with uni­ver­si­ties and research cen­tres, such as pre­lim­i­nary involve­ment with pro­cess­ing of mass data by look­ing at, for exam­ple, whether data are checked and eval­u­at­ed for unusu­al pat­terns. In this way, it is pos­si­ble to iden­ti­fy inter­na­tion­al net­works and whether tax loop­holes are being exploit­ed more rapidly.

Owing to the pan­dem­ic, com­pa­nies have been allot­ted finan­cial sup­port from the state, the VAT was low­ered for a cer­tain amount of time, and com­pa­nies had to adjust their cash­es as well as, in gen­er­al, their book­keep­ing.  All these activ­i­ties are asso­ci­at­ed with a lot of con­sult­ing, work of all kinds involv­ing time and effort. Do you think dig­i­tal­i­sa­tion will help your pro­fes­sion to react more quick­ly to such devel­op­ments, thus enabling a sort of press-the-but­ton adjustment?

Basi­cal­ly, the pan­dem­ic has again sig­nif­i­cant­ly increased the need for dig­i­tal­i­sa­tion and the focus on process­es. For exam­ple, in addi­tion to tem­porar­i­ly changed tax rates, a statu­to­ry mora­to­ri­um on pay­ments also meant that com­pre­hen­sive adjust­ments had to be made to the dun­ning process at short notice. Cus­tomers were afford­ed the right to sus­pend pay­ment owing to a legal mora­to­ri­um for a cer­tain peri­od of time which nat­u­ral­ly had an imme­di­ate impact on com­mu­ni­ca­tion and over­due notice pro­ce­dures. With regard to a bulk busi­ness, this means mak­ing a change in a high­ly auto­mat­ed process. These types of adjust­ments can­not be realised by press­ing a but­ton because, aside from a tech­ni­cal change, thor­ough prepa­ra­tion, assess­ment and also mon­i­tor­ing are required.

The pan­dem­ic has sig­nif­i­cant­ly accel­er­at­ed the need for trans­for­ma­tion and has also demon­strat­ed that the suc­cess­ful appli­ca­tion of new tech­nolo­gies makes it eas­i­er for com­pa­nies to react more quick­ly and to bet­ter man­age com­pa­nies even in chal­leng­ing times.

My key mes­sage is:

 

“The task of a finance organ­i­sa­tion is increas­ing­ly becom­ing that of an enabler for com­pa­nies in order to effi­cient­ly man­age and devel­op their busi­ness. The moti­va­tion to change through the use of new tech­nolo­gies and the open­ness to fur­ther devel­op team skills and col­lab­o­ra­tion struc­tures are key fac­tors for a suc­cess­ful transformation.”

Chris­t­ian Klein

Mr Klein, thank you for shar­ing your insights on recent devel­op­ments in the field of tax con­sult­ing and auditing.

Thank you, Dr Cal­daro­la, and I look for­ward to read­ing your upcom­ing inter­views with rec­og­nized experts, delv­ing even deep­er into this fas­ci­nat­ing topic.