Last week, the Finan­ci­al Times published an arti­cle, ‘Busi­nes­ses turn to soft­ware robots for office work’, that out­lined how, “Soft­ware robots have beco­me one of the hot­test fads in busi­ness auto­ma­ti­on, as a new wave of AI is poi­sed to sweep through the back-office func­tions of lar­ge cor­po­ra­ti­ons.”.

Various num­bers are thrown around in the article—a $153m invest­ment in UiPath and a valua­ti­on of $1.1bn for the same firm, Blue Prism valued at GBP 1bn with annu­al sales of GBP 24.5m, sub­scrip­ti­on reve­nue growth of 146 per cent for Auto­ma­ti­on Any­whe­re, an over­all mar­ket valua­ti­on of $2.9bn in 2022, and four mil­li­on workers in the US likely to see their roles auto­ma­ted by 2021.

First­ly, as we’ve writ­ten about befo­re, the repla­ce­ment of workers with robots will not be a blan­ket deve­lop­ment; the impact most­ly keen­ly felt in spe­ci­fic sec­tors, indus­tries, and roles. The arti­cle does cor­rect­ly point out that most of the workers will be kept on and retrai­ned to do dif­fe­rent roles. After all, the­re is always a mar­ket for the qua­li­ty work that only humans can do.

A striking par­al­lel that can be drawn is bet­ween fine dining and McDonald’s. The first McDonald’s restau­rant ope­ned in San Ber­nar­di­no, Cali­for­nia in 1940 but only began to grow into the mono­lith it is today from 1955 when its first fran­chise was ope­ned by Ray Kroc in Des Plai­nes, Illi­nois. The advent of chea­p­ly-pro­du­ced food avail­ab­le at low pri­ces did not spell the end of fine dining—high-end-restaurants still exist and do good busi­ness. Nor did people’s tas­tes change—in recent years, after deca­des, the public began to turn towards ethi­cal­ly-pro­du­ced, high-qua­li­ty food.

McDonald’s respon­ded to this by deve­lo­ping and ope­ning McCafe’s, aping the other end of the mar­ket. Auto­ma­ti­on will play out the same way—the public will pri­ze and value high-qua­li­ty, indi­vi­dua­li­sed work even more even­tual­ly. And that’s the remit only of humans.

But perhaps the most-pro­ble­ma­tic sta­tis­tic is also the article’s most-ano­dy­ne. “Each bot,” the report say, “can hand­le the work it would take three or four full-time workers to per­form.”

That num­ber, which ori­gi­na­tes for For­res­ter Rese­arch, seems incredi­b­ly low and makes one ques­ti­on if tho­se bots are being run on a Com­mo­do­re 64. At Ret­res­co, we can gene­ra­te indi­vi­du­al texts, ori­gi­na­ting from extre­me­ly-com­plex data, of around 12 sen­ten­ces at a rate of 24 texts per second. For one cli­ent, we rou­ti­nely gene­ra­te over 200,000 texts per day.

For­res­ter Research’s esti­ma­te of three-to-four full-time workers also for­gets that robots do not take breaks, or get sick, or go home at 18:00. Robots can work twen­ty-four hours a day, seven days a week, three-hund­red-and-six­ty-five days a year. And while robots do requi­re some main­ten­an­ce and upgrades, it is no more than what be expec­ted in any lar­ge orga­ni­sa­ti­on.

If you com­pa­re the pro­duc­tivi­ty of robots to humans, humans will always lose. A more-use­ful com­pa­ri­son would be to not look at which of the two sides is bet­ter, but how best they work toge­ther. The rela­ti­ons­hip bet­ween the two should be com­ple­men­ta­ry, not adver­s­a­ri­al.

* * * *

In a few days, we will post the second part of this ongo­ing series whe­re we will exami­ne prac­tical, real-world uses of NLG and Natu­ral Lan­guage Under­stan­ding (NLU) app­li­ca­ti­ons.