What can we learn from RPA failures?

The robot is no stranger to heavy indus­try, but its vir­tu­al co-work­er, robot­ic process automa­tion (RPA), is only just begin­ning to find a place with­in the indus­tri­al sec­tor. The technology’s clever soft­ware robots can ful­fil repet­i­tive and time-con­sum­ing tasks, and offer a mul­ti­tude of ben­e­fits from improved accu­ra­cy to cost-sav­ings.

How­ev­er, with RPA fail­ures a com­mon occur­rence in its ear­ly adop­tion, it’s clear that ini­tial imple­men­ta­tion of the tech­nol­o­gy has not proved to be smooth sail­ing for many busi­ness­es.

Why do so many RPA projects fail?

For multi­na­tion­al con­sul­tan­cy EY, RPA fail­ures are all too famil­iar, hav­ing wit­nessed 30 to 50 per cent of ini­tial projects fail. Com­pa­nies devel­op­ing the tech­nol­o­gy claim it can trans­form oper­a­tions, but if it’s as favourable as they say, why are there so many RPA fail­ures?

In actu­al fact, the ten­den­cy for RPA to fail is not down to the tech­nol­o­gy itself. This instead rep­re­sents a mis­use and mis­un­der­stand­ing of the tech­nol­o­gy.

“A fun­da­men­tal mis­take which is com­mon­ly made is that process automa­tion is seen as a pure tech­nol­o­gy imple­men­ta­tion deliv­ered by IT,” says Neil MacLean, part­ner at EY UK and Ire­land.

Oth­er com­mon errors that have led to RPA fail­ures include mis­judg­ing what hap­pens after adopt­ing automa­tion, think­ing robots are the whole solu­tion, direct­ing RPA tech­nol­o­gy at the wrong process­es and assum­ing the skills nec­es­sary for a pilot are up to scratch for real oper­a­tions.

Learning from RPA failures

The good news is that because most RPA fail­ures are the result of human error, rather than the tech­nol­o­gy itself, busi­ness­es can learn from their past mis­takes and sim­ply take a dif­fer­ent approach.

Aaron Bult­man, direc­tor of prod­uct at Nin­tex, which has its own RPA plat­form, says to avoid mak­ing the same mis­take twice, a busi­ness must map the entire process of adop­tion to ensure it has a full under­stand­ing. The com­pa­ny should involve sub­ject mat­ter experts through­out, so incor­rect assump­tions aren’t applied, and remem­ber that just because one process is not a good fit for RPA, it doesn’t mean the busi­ness should stop using the tech­nol­o­gy.

Where robotic process automation is transforming industry

RPA projects that suc­ceed teach us the same lessons as those which fail. And there’s a lot to learn from busi­ness­es that are thriv­ing as a result of imple­ment­ing the tech­nol­o­gy.

“While there are many use-cas­es of RPA in the office envi­ron­ment, where it real­ly comes into its own is in heavy indus­tries such as ship­ping, man­u­fac­tur­ing and min­ing,” says Johan Carstens, chief tech­nol­o­gy offi­cer of man­u­fac­tur­ing and auto­mo­tive at Fujit­su. “The biggest pri­or­i­ty in these indus­tries is work­er safe­ty and RPA is being used to reduce the depen­den­cy on humans in high-risk tasks.”

EY 2017

EY is work­ing with a min­ing com­pa­ny to devel­op a mod­el that helps its met­al­lur­gi­cal team increase ore recov­ery. The mod­el is based on math­e­mat­i­cal pat­tern recog­ni­tion tech­niques, or machine-learn­ing, and uses data automa­tions to make rec­om­men­da­tions.

Nasa uses Eggplant’s RPA plat­form to auto­mate var­i­ous process­es in its Ori­on space vehi­cle, which is being pre­pared for a mis­sion to Mars. “Equipped with three main dis­plays to mon­i­tor and con­trol the space­craft, to ensure the soft­ware behind the glass dis­plays oper­ates with­out faults, rig­or­ous auto­mat­ed test­ing is need­ed,” says Dr John Bates, chief exec­u­tive of Egg­plant. “RPA is sup­port­ing Nasa’s mis­sion to take humans deep­er into space.”

The cement man­u­fac­tur­ing indus­try uses machines to process a vari­ety of block prod­ucts and the con­crete mix used to gen­er­ate these blocks must be high­ly con­trolled. In the past, this process was com­plet­ed man­u­al­ly, but nowa­days RPA is used to open and close gates, mea­sure mois­ture lev­els, and reduce human inter­ac­tion to elim­i­nate error.

There are also exam­ples of RPA being suc­cess­ful­ly imple­ment­ed by com­pa­nies in the auto­mo­tive indus­try, such as Renault, as well as in the pro­duc­tion of radi­a­tion oncol­o­gy sys­tems, by com­pa­nies includ­ing Accu­ray, and with­in
sup­ply chain busi­ness­es.

By sav­ing mon­ey, reduc­ing human error and min­imis­ing repet­i­tive tasks to free up time for employ­ees, RPA has the capac­i­ty to trans­form heavy indus­try. But busi­ness­es must apply the right meth­ods of adop­tion to ensure they make the most out of the soft­ware and do not leave stake­hold­ers won­der­ing why they invest­ed in the tech­nol­o­gy in the first place.