Is RPA officially the new outsourcing?

May 17, 2018 | Phil Fersht

Just as many enterprises were running out of places to find more and more hidden costs they could quickly remedy through (yet) more outsourcing, along came their perfect new toy to unearth costs they had never thought possible to eliminate: RPA.  

Yes, folks, this stuff is just the thing to keep you occupied for the next few years to keep your greedy CFOs at bay - and even includes the word "robot" to conjure up images of human work displacement, creating hours upon hours of repetitive (robotic) debate at conferences from people who literally sprung from seemingly nowhere to become lifelong experts in this new dark art. 

And, oddly, most of these new RPA maestros seem to be exactly the same people who were hawking the delights of outsourcing just a couple of years ago.  Maybe the connection between outsourcing and RPA is a lot closer than we think?  So let's have a gander at the new findings from the 2018 State of Operations and Outsourcing study, conducted with KPMG across 381 Global 2000 organizations, where we questioned operations leaders about their intentions to keep investing in RPA and outsourcing. 

This data shows the tranche of operations leaders making significant investments in RPA and outsourcing, sliced by industry sector:

Financial services firms, where outsourcing is most mature, are showing voracious appetites to go down the RPA path

While banks and insurers are showing the smallest appetite (10%) to keep pursuing aggressive outsourcing strategies, they are right at the front of the queue (50%) when it comes to RPA.  Insurers were one of the first industries to explore BPO and offshoring twenty years ago, so it's little surprise that RPA is so appealing to these firms, where they can find completely new ways to mimic highly repetitive, intensive processes, plagued by manual workarounds, using smart software solutions.  In addition, many of these firms have been outsourcing back-end processes that have become very stale over the years, and RPA provides the perfect shakeup to rethink how to rework these. 

Sometimes RPA provides the perfect catalyst to force an outsourcer to get back to the table to reinvest in their client or risk losing them to a hungrier competitor. Banks have always been a bit weird when it comes to outsourcing - they have tended to move massive amounts of IT development and maintenance work to service providers over the years, in addition to infrastructure, but have been very shy when it comes to BPO, often preferring to move process work into their offshore shared service centers, citing issues around privacy and compliance as their reason to keep it inhouse. It's surprising that the appetite to explore RPA is so strong (58% making significant investments this year) when you consider that most banks have to comply with various regulations which necessitate a human to oversee pretty much every process that is conducted within their organization. However, with the sheer quantity of legacy detritus plaguing banking IT systems, such as spaghetti code that began its life several generations of programming languages ago, where some of the original cobol guys who started it have since deceased (no joke), and mainframes that really should be moved to one of Kim Jong's testing sites, RPA can actually help breath new life into fixing some of these processes in a way that can have a massive transformative effect on their operations.  (Read our POVs on Banking and Financial Services RPA uptake here to learn what 80 of them are doing, and read here to deep dive into the insurance sector and its attitudes towards automation.

Industries that just need to shed costs as fast as they can to remain viable are aggressively jumping in

Telecom was always a bit late to the game when it came to heavy outsourcing, partially because its systems are so complex and they are so dependent on microtransactions which are very difficult to outsource.  However, the high throughput, high-intensive nature of telecom processes places the industry right at the forefront of RPA appetite.  With such a strong impetus also on outsourcing, expect to see more of these automation-led outsourcing deals transpire.  Utilities firms, on the other hand, still tend to be very slow adopters of new models, and most are still very focused on getting their outsourcing models operational, after many painful years of dealing with labor unions and archaic IT systems.  Surely RPA beckons soon, but expect this sector to be behind the others.

Retailers have always struggled from decentralization (often growing through many years of painful M&A) and horrific ERP experiences.  With the pressure to adopt digital customer channels more intense than ever, RPA does provide some significant benefits to fixing legacy processes that were simply not cost-effective to outsource in the past.  It's a similar story for travel firms, especially those making major efforts to up their customer digital experience.  RPA can be a huge help linking customer portals with back-end systems that have suffered from manual workarounds and poor integration for decades.

Manufacturers have been one of the pioneers of outsourcing, especially as many focused on core supply chain areas first, before moving onto IT and BPO in more recent years.  Most manufacturers ran out of room to optimize their outsourcing engagements many years ago, and stagnated when it came to improving poorly-integrated supply chain and accounting systems.  These firms are all about driving out every ounce of cost, and if they can do it, while making process fixes they have neglected since the days of MRP and JIT, then RPA is something they really want to get moving on.

Energy firms have always been massive outsourcing customers with a strong SAP underpinning - both for IT and BPO work.  With the massive cost pressures impacting energy firms, and a major impetus to transform their operating structures away from legacy labor-intensive models, it's no surprise that energy firms are among the early adopters of RPA (click here to read more about the transformation issues facing oil and gas firms, and here for a decent case study of NPower and its RPA experiences). 

Healthcare has always been the "odd duck" when it comes to operating models and transformation.  Starved of funding, held hostage by unions, and a culture of never changing anything, healthcare is typically at the back of the line when it comes to being bold and exploring radical new opportunities like RPA. However, with the tumult being created by the impact of Obamacare and whatever is going on with its unraveling, there are pockets of healthcare organizations now exploring more innovative ways of saving themselves, and RPA can be very effective for many (read here for some greats example of how some healthcare orgs have adopted RPA).

Bottom-line: As much as we hate to admit it, this is looking like the new outsourcing.  But the longer term impact is very different....

However which way we were looking at it, the outsourcing space was slowing down, and it's hard to get too excited about a market growing at 1-4% each year.  We have been brought up in a world promising 50% savings, and achieving 30% (at least on paper).  We needed to find the next thing to grab onto that was back-office focused, a bit messy, quirky, and loaded with hype.  I could bore you to tears with a lost of caveats of how to avoid screwing up your RPA, how to focus on "value" and not "cost" (who are we kidding), how you need to align business and IT, how you need to get right on top of change management and cultural impact.  But read the RPA Bible if you want all the caveats, the best practices, the pitfalls... and how to avoid RPA hell

For now, I think this is the "new outsourcing", where deals are spiked with RPA to deliver the numbers.  So time to love with what we have created, and see if we can somehow make it all work... 

Posted in: Business Process Outsourcing (BPO)IT Outsourcing / IT ServicesRobotic Process Automation

Never Miss A Story

Sign up for the HfS Research newsletter and get the best research delivered to your inbox weekly.




  1. Stan
    Posted May 10, 2018 04:16 PM | Permalink Reply

    Spot on and I don't hate to admit it, but rather view it as it is a good, natural and logical next step in the evolution of service delivery and more practical and easier (though not easy) to successfully pull-off than mass human based outsourcing. Also, efficiency and to a lesser degree effectiveness gains albeit with job loss via automation has much more historical precedence and learnings from which to draw and also instills only a general sense of glum resign among impacted workers as opposed to virulence spawned by moving jobs "offshore." While RPA is not as easy as advertised (i.e., ideally involves upfront process streamlining work, is not a one and done, only eliminates portions of jobs, requires solid change management and governance capabilities) it can ideally enable the long sought goal of focusing human talent (assuming such talent is available and has required adequate skills and resources) on more value-added and strategic work as opposed to rote transactional work that typically only has a major organizational impact when done wrong or outside the bounds of various regulatory and compliance dictates. So while we will all have to suffer new-age RPA luminaries recently repurposed from being outsourcing pundits not long after being business process re-engineering savants, the gain in total to the services industry and all user organizations struggling with the twin conflicting mandates of becoming more strategic while still saving more money will benefit.

  2. Mohit Verghese
    Posted May 10, 2018 05:06 PM | Permalink Reply

    Phil this has already happened, lot of companies now provide transaction price seperate for humans/bots

  3. Ralph Diaz
    Posted May 11, 2018 04:57 AM | Permalink Reply

    Interesting. A question now Phil: will (and should) RPA (development and operations) be outsourced to existing outsourcers?

  4. Sunder Sarangan
    Posted May 10, 2018 11:08 PM | Permalink Reply

    Only an analyst can answer this question correctly, but if RPA is indeed the new outsourcing - then arguably AI is also the new computing.

  5. Phil Fersht
    Posted May 11, 2018 05:02 AM | Permalink Reply

    Spot on Sunder. everything is pretty much the same as previously - we sell boxes, developers and bullsh*t all packaged up in different flavors - and paid for via an as-a-service model. And - heaven forbid - we dare to make any changes to the process, then we pay the expensive consulting fees and off we all go again =)


  6. Partha
    Posted May 10, 2018 10:14 PM | Permalink Reply

    Hi Phil: Agree that its the new wave and focus area for many industry verticals. But in my humble opinion, I feel that the enterprise leaders polled didn't interpret the question correctly. While RPA once implemented, will eliminate manual BPO efforts, the execution model for these RPA solution implementation programs today is pre-dominantly offshore-onsite ( aka "outsourced"). So in essence if their intention is to spend on RPA, they are indirectly spending on outsourcing as well?


  7. Phil Fersht
    Posted May 11, 2018 05:11 AM | Permalink Reply

    @Partha: This wasn't a single question. I took their outsourcing intentions (separate question) and compared with their RPA investment intentions. This is why the data is so startling, we were not asking them to compare apples with oranges (or whatever that phrase is). What is clear is that they are looking much more aggressively at RPA as their cost containment vehicle, than they are outsourcing:

    What is happening very quickly is most the service providers are already combining bots and humans within their bids (as Mohit Verghese points out). Hence Rick Vargas' "botsourcing" phrase is quite accurate!


  8. Rick Vargas
    Posted May 10, 2018 08:03 PM | Permalink Reply


  9. Jennifer Valenti
    Posted May 10, 2018 06:40 PM | Permalink Reply


  10. Phil Fersht
    Posted May 11, 2018 05:49 AM | Permalink Reply

    @Ralph -- They are already trying it. I am having a tweet conversation with @Jaako_Lehtinen at Capgemini about their "Intelligent Automation as a Service" offering. You should check it out - Cap hosts the bots and has an on-demand service to support them. Any "change to the process" is an immediate consulting order, but am not getting too much clarity regarding what happens to the hosted bots when the client cancels. But the model does sound interesting, provided the client has its act together when it signs up. IBM is aggressively pushing its automation services capabilities and has won some large deals in this area:

    Accenture acquired GenFour (the only pureplay RPA managed services of some scale):

    And we're seeing a ton of RPA noise from the likes of Wipro, Hexaware, TCS, HCL and others.

    One of our analysts, Elena Christopher, is leading a Blueprint report that covers RPA services in-depth, that includes all the major service providers, in addition to consultants and boutique pureplays. Will be interesting to see ISG competing for the same RPA consulting business as the service providers!


  11. Francis Carden
    Posted May 11, 2018 06:58 AM | Permalink Reply

    I sold my first outsurrcer RPA deal in 2007 and many more since. But I got frustrated that BPOs, who were on such tiny margins weren't even happy if we gave them them RPA for 1/10th the list price. So we went to the customer directly. Our first telco contact center was automated with 35,000 RPA attended bots which were then forced on their 6 BPOs.

    Businesses are not stupid. They are not going to pay an outsourcer (for long) if it's proven easy to automate with RPA. The BPO Guinea pigs are losing out because they ignored RPA when they could have taken the field.

    Having said that, outsourcers don't often get access to the actual systems which limits them to VDI and OCR screen scraping. Making RPA complex (and more expensive than the customer doing in house) . Not impossible but complex and expensive given the much slimmer ROI). If the enterprise automate the easy stuff quickly, the BPO constantly gets the most complex stuff, then what's left, gets more and difficult to automate. Not just complex processes but fragmented and down right dirty/complex apps.

    I can share many use cases where BPOs are using our RPA but where the robots were built by our customers. That one that did this in 2007, those BPOs are still using today, whether they want to or not..

  12. Tony Bocock
    Posted May 11, 2018 08:06 AM | Permalink Reply

    I find the parallel RPA / outsourcing rather irrelevant, based thinly on the idea that instead of outsourcing to people, firms outsource to machines. The fact that more money is being spend on RPA than outsourcing is however very interesting. I would have presented the comparison as "one of many" technological innovations that firms are implementing to improve their levels of performance and efficiency. The other items in the "many" basket include IoT, Big Data, IA, Blockchain... Firms now have a much wider range of options to improve their performance - all based on outcome from the digital revolution. I would argue that outsourcing has been just one of the (earliest) of those options - based on the technology of large internet bandwidth enabling the tasks to be done outside of company offices and from there offshore. Thanks for the figures !

  13. Ed Susman
    Posted May 11, 2018 11:53 AM | Permalink Reply

    Phil, I think the key phrase in your conclusion is "for now". My sense is that RPA is a transitional step in the move from human labor to digital labor. As such RPA is replacing humans in the sourcing equation and becomes the new sourcing paradigm (in or out...take your pick). The question is what will happen as we continue the paradigm shift. Does Intelligent Automation become the new outsourcing, replacing RPA or do we evolve to a completely different model where what has been sourced no longer exists.

  14. Phil Fersht
    Posted May 11, 2018 12:01 PM | Permalink Reply

    @Tony Bocock -- agree that the models between legacy outsourcing and botsourcing (thanks Rick) are very different. But the common thread is the broader outcome: these companies need theor next threshold of efficiency and effectiveness and RPA has somehow bulldozed itself into that place. Yes, we can bemoan how hard this is, and how companies need to invest in some real process and change management capability, but the bottom line is they need something to hang onto. Yes, you can move transactional work to a bot. Yes, you need to have your act together to do it. Yes, you will probably screw up a few times trying and waste a ton of money in the process. But RPA is the one thing providing some hope that we haven't reached that boring place where we have run out of magic bullets. We run out of magic bullets and we will end up like the telecoms business of the early 2000's...


  15. Phil Fersht
    Posted May 11, 2018 12:06 PM | Permalink Reply

    @Francis -- BPOs hit a wall because they were not being let in behind the lovely citrix firewall. You can't do RPA is you can't get more intimate data access. Hence RPA is a conduit to take what was left of BPO into a new era of sourcing.

    Now will all these BPOs make the investments to get there? Of course not, but some will... I expect half of them will shrink or get acquired over the next 2-3 years and we'll end up with an industry of tech-driven outsourcing players operating at a bots/humans level we never quite anticipated.

    However which way we look at this, the BPO business needs a catalyst to break through the lethargy. Hopefully this will be it...


  16. Frank Pendle
    Posted May 11, 2018 01:45 PM | Permalink Reply

    I think we're at the point that outsourcing and automation are converging. I expect large-scale RPA deals to start "popping" as companies start realizing large-sale integration (hat tip to OneOffice) and the SIs themselves start looking for the next large wave of efficiencies promised in their multiyear BPO SLAs.

  17. Yogi Singh
    Posted May 11, 2018 12:12 PM | Permalink Reply

    RPA is a super hit only where there are high # of business transactions and legacy systems (also mainframes). Insurance/BFS are classic example of both. The other verticals are quicker in digitising their legacy systems. RPA is just a stop gap arrangement before BAPI and intelligent networks take over. Outsourcing will survive for simple reason: when market slows down, you down size and look for arbitrage benefits, and when you grow fast, you need help in on-boarding new resources (better still in outsourced model). BPO is maybe a shock proof industry and will probably continue to grow at a steady pace for a long time.

    There will always be Creators and Processors. Creators will always avoid becoming Processors. Creators always want to be up skilled/grow business/find newer pastures and Processors will just like to pick up and run with whatever Creators get bored with. Creators are IT/Business and Processors are your BPO guys.

    Processors will pickup these boring tasks, execute them faster, better, cheaper and simplify them to their deathbed.

    I wanted to call Processors as Executioners originally.

  18. Francis Carden
    | Posted May 11, 2018 04:15 PM | Permalink Reply

    I have a great example of a large tech company implementing a new digital CRM back in 2010. Very cool tech. However, they used 20,000 RPA attended bots for just over 3 years to plug the gap whilst they built their truly digital back-end APIs. then retired the bots. The automations were deployed in 6-9 months. The cost of building and managing RPA unattended bots only works if you get into the element of enough scale. Too little and/or too hard, there's no point. Too much (easy or hard) and likely the cost of replacing the bots with true digital transformation is easily viable - and/or already started and like the first sentence, just becomes a stop-gap.

    I only say this because the technology of RPA is not new to enterprises, BPO's or IT. (I've been selling it since 2005, and perhaps even long before). What's new is the term Robotic (scraping) and whilst there is definitely value as it's become very mainstream, your previous points of scale are becoming more and more evident.

    My new point to enterprises is try to push the end of Legacy strangulation. Wrapping Excel with RPA sounds good except when Excel is used to wrap MS Access, which imports data from SAP. Companies will die by this slow strangulation if they don't also transform. The BPO's rarely help with transformation but the SI's surely know how to.

  19. Aman Mustafa
    Posted May 12, 2018 08:29 PM | Permalink Reply

    Outsourcing, in my opinion, is there to stay for a long time since the user-companies are not going to change their annual IT projects dramatically down so it can be done internally - also due to the fact that internal resources are always not up-to-speed on new technologies, viz., RPO, AI, ML, CC, Blockchain, etc. Hence, they would rather outsource the companies that are capable of delivering such technologies. Somebody has to write the code for the RPAs to function and change as the usecase changes. Hence my first line of opinion. -aman

  20. Rudi
    Posted May 14, 2018 12:39 PM | Permalink Reply

    Interesting and to-the-point as always Phil. This said, I'm also wary that the same strong adopters of the outsourcing model are now queuing to be at the forefront of RPA projects, while still not doing the all-important strategy exercise first. This time hopefully not, but probably again being disappointed by its outcomes after having spent quite a bit of the shareholders' money on lustrous initiatives.

  21. Terence Fogarty
    Posted May 17, 2018 11:41 AM | Permalink Reply

    Excellent perspective. Thank you for sharing. The analogy is perfect IMHO.

Post a Comment

Your email is never published nor shared.