RPA hype versus reality - an early look at use cases and data from HfS FORA NYC
- Summary:
- Robotic Process Automation is a potent topic because companies are far enough along to share proof points - and frustrations. But is RPA just an efficiency driver? And how does AI fit into the mix?
Den Howlett surprised me with his optimism about Robotic Process Automation (RPA) at the HfS FORA UK event (see his roundup: FORA Summit 2017 paves the way for RPA success). I've been looking forward to a firsthand look at HfS FORA New York City ever since.
"We're past the trough of RPA disillusionment"
Now on day two, we are just getting into the depths of RPA talk (day one dove into a broader context of industry change, AI, machine learning, blockchain and so on). I haven't heard a use case as compelling as Den's HSBC example from the UK yet, but we're now in the midst of an enlightening RPA debate. A panel of RPA practitioners and service leaders noted the following:
- many companies are still struggling to get automation to work
- there is an issue with RPA sustainability as tech and processes change
As one panelist put it, we're past the trough of RPA disillusionment into acceptance of automation as valid tactic. But:
70 percent of enterprises don't know how to spell RPA just yet.
I'd add:
- there are issues with RPA scale (most enterprises don't have an RPA center of excellence for example)
- more published use cases are needed (more on this shortly)
But the panel also agreed: RPA laggards are going to get hammered:
RPA penetration is "decimal points versus the opportunity" (by decimal points, we're talking .1 percent)
Industry leaders are plowing ahead:
good line - "If you think automation is hype, you shouldn't own any Google stock....." #hfsfora
— Jon Reed (@jonerp) March 8, 2018
HfS Research Founder/CEO/Chief Analyst Phil Fersht kicked off the panel with this eye-opening slide:
The slide, from a recent study by HfS and KPMG, shows that RPA outpaces cloud, IoT, and analytic as areas of "significant" enterprise investment (RPA is the only "significant investment" that scored above 50 percent).
Learning from RPA use cases
With this type of investment underway, buyers need use cases to ponder, and peers to learn from. We got that on HfS FORA day one, via a talk from Tony Saldanha, VP, IT and Global Business Services at Procter & Gamble. In his keynote, "The next generation of share services: how to successfully transform your GBS operation into OneOffice," Saldanha walked us through an RPA framework before sharing a few P&G examples.
If you're wondering about OneOffice, Den covered that off in his piece, but it's an HfS-branded concept on a new way of thinking about operations - without a front/back office barrier. P&G has a number of RPA-type projects under their belt. Saldanha touched on A/Rex, or "Accounts Receivable Exponential," which automates deductions and claims processing. Saldanha:
Before we built A/Rex, we used to have a couple of hundred people that looked at deductions and claims that came through from our customers, like "Hey, you said you shipped us eighty trucks, we just got seventy-nine," and they'd have to determine whether we'd have to refund some money.
The A/Rex algorithm has changed that:
For 50 percent of the cases, it can determine with 99.9 percent accuracy whether the claim is valid or not.
As for the other 50 percent? A/Rex is still 94 percent accurate. But Saldanha was quick to point out: this doesn't necessarily mean headcount reduction.
You don't have to give pink slips to half your organization. What you can do is use that same capacity to look at other receivables that you should be claiming money back from customers that you paid inadvertently. There's always ways to make this a win-win thing. It does not have to be a human win-lose kind of situation.
In keeping with the OneOffice theme, Saldanha walked us through Project FOX, or Front OfficeX. Front OfficeX is driven by a provocative question:
The whole idea is in today's day and age, why would any company, as part of their shared services offshore or onshore center, ever have a front office? Why would that just simply not be self-serve?
The first Front OfficeX use case was the finance organization. There are about 500 full time employees (FTEs) serving as the front office for the financial operations of the parent company. They support 25,000 P&G employees:
They have questions ranging from, "I'm a branch manager, and I'm going to launch a new product, to, "I need a budget unit code, can you help me?" [They also get questions about] closing the books, or accounts receivables, or accounts payables. They get about 1,000,000 requests annually.
That's a big call center burden:
We said, what if we use chatbot and voice to essentially have those same associate brand managers do their own stuff?
This system is now live, and automates the creation of an internal order. One big payoff? Training is easy:
In your consumer lives, you guys have done something similar, so we never had to train anybody.
The time saved from old clunky processes is dramatic:
In the previous process, you either filled in a form, or you called somebody, and then you gave them information. Then they took two days to go into the back end system, make those upgrades or changes, and then come back and say, "Here's the answer."
The chatbot gets "smarter" over time:
As this chatbot executes, it actually learns. It knows which process tree it's following. As it gets more hits on certain types of questions than others, it flags that for you... so you can actually make this more tailored to the kind of questions you're getting. That's in addition to the machine learning that it has to actually improve the answers.
Front OfficeX is live, and, as per Saldanha, "It's delivering value in terms of dollars and obviously the agility of doing business, and things like that."
My take - from brute force to intelligent automation
If you're looking to make a case for front office automation, Saldanha's numbers wet the RPA whistle:
- 40 percent reduction in contact volume via bots
- 1/100 per transaction versus humans
- 84 percent of users satisfied with the self-service experience
But during the Q/A, the audience came back to the question of job loss. I spoke with several customers looking at RPA, all of whom emphasized efficiency, which could lead to head count reductions. But that's not how Saldanha sees it:
The idea is obviously not to eliminate all administrator jobs, it's always to keep moving people to higher [level] jobs.
He added:
Saldanha of P&G: 5% of any process you do, you could use RPA and get rid of that 5% (that % is growing by 5 every year)
At the high end, 13% is human work you couldn't replace"The vast middle" is where you can use automation to enhance human work e.g. call centers #HfSFORAz
— Jon Reed (@jonerp) March 7, 2018
Yes, companies are adopting RPA, but I don't see many examples of transformation at scale:
This talk on RPA is hitting the missing link in terms of scale. Instead of "this changes everything" blind tech faith I am hearing:
Set up an RPA COE (center of excellence), and make sure the owner/director can bridge convos between biz and IT. Now we're talkin'. #HfSFORA
— Jon Reed (@jonerp) March 7, 2018
— Jon Reed (@jonerp) March 7, 2018
I'm now in an "intelligent automation" breakout session, where I hope to learn more about how buyers are using automation in an elegant - and not a brute force - way. I still see too many "brute force algorithms" in play, a topic I slammed in Algorithms are undermining the customer experience, and (most) companies don't care.
Finally someone (@chirag_mehta) points out that automation is often misapplied with brute force automation of a call center - yes, it was efficient,
"but it created a miserable user experience"
"let's not repeat the same mistake" - #hfsfora
— Jon Reed (@jonerp) March 8, 2018
My frustration with unintelligent automation lines up with another HfS Research stat: 80 percent of companies have yet to connect AI capabilities with their RPA systems. So my quest for intelligent automation continues. I'll report back after day two - as buyers share their highs and lows.