Refining the robotics argument for finance

Den Howlett Profile picture for user gonzodaddy September 15, 2015
Robotics for finance is stuck. The models of BPO and IT are poorly aligned and this is contributing to dis-ease at the prospects for improved processes.

robots in tech
Chris Middleton's dissection of an ACCA report The robots are coming? Implications for shared services (PDF) was an interesting discussion that made pertinent points but from flawed assumptions. Like Middleton, I believe the underpinning thesis for the ACCA report is troubled. It starts:

Finance leaders obsess about transformation levers – people, process and technology. But today’s solutions, save the requisite investment in an expensive ERP platform, have typically leaned heavily on the equation of people and process – cheaper people located offshore delivering transactions, plus added process improvement equals significant savings. The technology component has mainly been limited to communication widgets that facilitate workflow and e-invoicing.

Enter the robots. A very clever marketing tool, robots, robotics, robotics process automation, applied automation – whatever you call it – conjures up an image of a machine replicating the activities of a human doing the work. It is evocative, it’s high tech, but most importantly, it’s emblematic of what some see as the next step in the evolution of business process delivery – fewer people in favour of intuitive, machine-based learning technologies.

The opening paragraph relegates the technology element of shared services to that of technology only - 'communication widgets' tells you all you need to know. This is wrong.

Field reports among enlightened technology and BPO operators suggest that integrating operations and IT but with the end user outcome at top of mind opens the door for significant efficiencies, the reduction of duplicated effort and streamlining of services that can readily accommodate a degree of robotic process automation (RPA).

The language used in the second paragraph is pejorative.  'Very clever marketing tool' and 'whatever you call it' suggests a jaundiced view. Unfortunately that's not stated as clear bias but is unsurprising.

State of the art

Where I disagree with Middleton is when he says:

First, the lines between human judgement and automation are already blurred, particularly when it comes to that most compliance-heavy sector and business function, finance...because of the rising numbers of organisations that instruct their employees to behave like machines, and thus remove the potential for human intervention. In accounting and finance, human intervention can be of paramount importance.

His opening statement is based upon an incomplete understanding of what happens in the accounting profession and in accounting departments.

The UK system of regulatory reporting is based upon making judgments. However, in an ideal world, accountants want to narrow the opportunities for making judgments because judgments are prone to error. That's why auditors have developed sophisticated systems of statistical sampling.

Those audit systems are designed to ensure that the amount of manual and detailed testing is focused on the most likely areas of significant risk. Most of the time that method works well but requires a good understanding of the systems in use. So the modern focus is on systems testing rather than transaction 'tick and bash.'

On the routine handling of the transaction, Middleton is right to assume that skilled human intervention is paramount but that is not a contrarian argument against RPA. Rather, it is an argument for asking how to solve the broken processes that lead to error and rework in the first place.

Frustration in the ranks

HfS end user study
HfS end user study

That is the central if understated thesis behind Horses for Sources call for RPA in a re-imagined finance function. In a recent polemic Phil Fersht, HfS CEO said:

I don’t believe I’ve ever witnessed a time our services “industry” has descended into such a mind-numbing pattern of meaningless marketing hype and unintelligible bullsh*t.  It was bad enough when everyone was painting pretty pictures of “transformation” to sugar-frost labor arbitrage deals, but at least we could understand the bullsh*t and detect its scent a mile away.

Frustrated? You bet and that is reflected in the lukewarm responses I read in the ACCA report. Middleton is right to think that accounting types are cautious but if the business cases put forward are weak then I'm not surprised that ACCA concludes:

The adoption of robotics in finance delivery is challenging. Finance leaders will only adopt RPA when they see peers they trust implementing these solutions.

CFO functinoal cost finance function cost analysis

Stasis is not compulsory

This is not a topic that's going away. Last month, produced a useful set of analyses covering where money goes in the finance function and the possible place of robotics: is striking that 62 cents of every dollar spent today to operate the finance group is spent on people...About one half of the typical finance function effort is devoted to transaction processing such as payables or remittance processing.

...all this goes to show that CFOs are skeptical about claims that process automation is the cure-all. Surely, one can engage in process streamlining and aggressive workflow automation. But on the other side of those commercial transactions sit suppliers, customers, and even employees. Until software robots can handle the sensitivity of interpersonal communications with grace and aplomb, we still need people.

That aligns closely with some of ACCA's findings and Middleton's argument about the dehumanizing impact of technology. But again, the argument ignores potential upside, and suggests a continuance of the status quo. That cannot be right.

Looking forward

Hardly a day goes by that some vendor or other is pushing analytics in my direction. Only yesterday I took a briefing on how operational modelling can be slotted into performance management systems normally run by the finance department. That is serous value add, representing activities where finance can be of real value. Where will those skills come from if finance is devoted to a focus on transactional accuracy and other routine tasks?

I'd like to think we are approaching a point where the finance professional wants to be persuaded. As I read the various reports and analysis, it seems that finance is saddled with competing interests from outsourced IT and BPO that have never been properly aligned but which are blinded by SLAs that keep turning 'green' and for which everyone pats themselves on the back.

Final words

Middleton poses a question about how the accountants of tomorrow will gain the right experience in a world where RPA is pervasive. I was brought up in the profession at a time when auditors' ink was specially made and locked away at lunchtimes. The concept of marginal costing was a relatively new science and we learned about standard deviation with only the aid of rudimentary calculators. I can assure everyone that I would not wish the 'old' days on any trainee! We live in far more interesting times.

A grey colored placeholder image