FORA Summit UK 2017 – the robotic process automation (RPA) hype is over; now is the time to get stuff done


RPA is rapidly moving from hype to action and what I heard and saw gives me hope that for once, we’re seeing technology that can easily deliver more than a point improvement


Last week I attended the UK leg of the Future of Operations in the Robotic Age (FORA) event organized by HfS Research. Beyond the fact HfS CEO Phil Fersht and I go back more than 20 years, I’ve been sitting on the sidelines waiting to see what FORA would produce. I was not disappointed. I’d go so far as to say it was the best (and last) event I attended in 2017.

How do I gauge that? I sat through every presentation, learned a great deal and enjoyed some post-event conversation with other analysts, practitioners, and consultants. That almost never happens at other events.

Logistically, Fersht told me they could have filled the room three times over. I don’t doubt it. While this audience only numbered 120 or so, it was a standing room only event. That’s rare at a time when executives are spoiled for choice during the events season.

With that bit of color out of the way, this event differentiated by offering a combination of solid content from analysts, practitioners, and customers, all delivered in a no-nonsense way. In that sense, it was a typical ‘British’ affair, with minimum fluff and barely a hint of ‘flogging’ on the part of the promoting event partners.

Sideline banter

It helps that British people tend to be forthright so, for example, it was fun to learn that Accenture has and is continuing to cannibalize itself while positioning itself as a genuine world-class consulting and delivery organization. We admire what Accenture is doing in the market, coming as it does with crisp messaging that makes sense.

Equal fun came from the lead architect of a well-known building society cum bank who complained bitterly that microservices are only SOA revisited (please let’s not go there.) And then there was the recently acquired software vendor attempting to make the case for additional fees for software if it turns out the software delivers more value than was originally intended. He openly made the argument for increases in the maintenance fee take, something that no doubt will come under greater scrutiny in 2018.

Those conversations aside though, if this event is indicative of where we are at, then RPA is rapidly coming of age and delivering genuine value.

The RPA/OneOffice rationale

Fersht kicked off the day with the analysts’ usual tsunami of PowerPoint driven charts – in this case, an 81 slide deck. While these usually send me to sleep, they acted as a good prop for telling a story that can be summed up this way:

Businesses are realizing that continued prosperity means getting away from so-so customer care and attention. It really is a time when customer service matters as evidenced by what Amazon, Netflix, Apple, Airbnb, Uber, and others have brought to the customer experience table. For many businesses, aligning to customer need is only possible within the context of constrained budgets. Therefore, anything that frees up the needed resource so that companies can staff and upskill to meet customer need gets attention. RPA provides a fertile ground to meet that objective, offering multiple opportunities to deliver what BPR promised in the 1990s but never really delivered. Will this time be different?

The answer is a qualified ‘yes.’ HfS believes that BPOs are the obvious providers, but what was clear from all participants is that this is not simply a labor arbitrage issue but a software and skills issue. Similarly, it is not a matter of deploying a few or a few hundred robots into business processes and then hoping for the best. It is a methodical examination of those areas where inefficiency can be carved out, starting with the big obvious wins but again in context.

HSBC – customers first – for a change

Recounting an inbound call center incident that he says is becoming typical, Warren Buckley, global head of contact centers, HSBC said:

There is a sense in which we have forgotten that we are meant to be providing a human to human interaction as the natural way we expect to get things done. So I don’t want to put robots everywhere to replace the human but I want to have tools that help the call center agent do a better job.

On the fundamental changes he sees, Buckley talked about how the rise in mobile use is leading to new behaviors. Customers are no longer seeing that they may be using a credit card when making a purchase via Facebook acting as an intermediary for PayPal which is tied to a credit card, even if the transaction turns up on their statement as a credit card entry. Similarly, he notes that ccustomers want to be served across a variety of platforms and not just the business website or app. Finally, he talked about how the time of contact has changed with spikes occurring on Sundays, early mornings and evenings. Which leads to a situation where:

I cannot continue to meet customer needs if I am solely relying on front-line customer staff. It’s not even a cost thing. It is getting harder and harder to hire people into front-line operational roles. Therefore, for me, robotics is an imperative. Anyone who is not considering robots for a variety of tasks is missing the point. To frame this, my demand in volume is falling 12-15% per year. My demand is growing by 20-25%. The complexity of calls I’m receiving is increasing.

HSBC sees robots helping in funneling customer requirements. this is achieved by training robots using the extensive library of knowledge management tools that HSBC has built up. Interestingly, in order to ensure acceptable quality standards, HSBC is using machine learning to augment the teaching of robots across multiple languages. In the call center, HSBC is augmenting the call agent systems via ‘active listening’ which in turn pops up information the agent needs as she works through the customer call.

We do this because I want the agent to be really good with humans and not really good with systems.

At the same time, HSBC has used the RPA initiative to start to consolidating agent environments, making it much easier for agents to get the job done rather than having to wade through screens before they can get started.

Another piece of the bigger picture comes in making the management of the global 20,000 agent workforce more efficient through the use of machine to machine processing and RPA. That has led HSBC to examine the role of data generated by the contact center, which in turn means HSBC has moved half the people working on spreadsheets over to work on developing customer insights.

Do I think robots will take over in the next year or even two years? No. I fully expect that within two years, everything we do will support customers via digital channels. This time, instead of focusing on cost and efficiency, actually pretty much everything except the customer, we have to focus primarily on the customer and the agents that support them. It gives us an iterative opportunity to learn. I think about 50% of that will come through some sort of robotic layer.

HSBC is what HfS would likely call a solid example of a company that is making its way towards the OneOffice concept that HfS has defined. OneOffice is an attempt to rethink how businesses are designed such that the lines between the front and back office functions are no longer important but supportive of one another with RPA acting as a kind of glue that not only streamlines processes but provides a way to radically improve the way in which customers are served. Phil Wainewright would likely characterize this as XaaS.

Fast pace

Coincidentally, Vinnie Mirchandani had this to say  in an article primarily aimed at reeling in AI expectations:

Far more tactical implementations of Robotic Process Automation (RPA) are more promising. RPA is not machine learning, it uses software bots to mimic human activity – like logging in to a system, copying and pasting data across systems. The payback is much lower than what AI/ML promises, but it removes the drudgery and increases productivity of many white collar jobs.

Mihir Shukla of RPA vendor, Automation Anywhere told me in an interview for Silicon Collar

“Think of the impact of this on real estate requirements. I can take what would need offices in six large buildings and put them on two racks. I can help consolidate global offices. A cognitive bot can process an invoice in 30 languages—I don’t need centers in the Philippines and other centers in Japan to deliver diverse language skills. Think of the employee recruiting and onboarding time we can help save.”

It is a more traditional definition of savings that comes with automation, and far more realistic. But even there, many services firms are pushing RPA which is a bit of a red herring when it comes to implementation costs.

It is a good indication of how quickly an industry segment can move that Mirchandani’s story while pivoting off the back of a very recently published book and making excellent points, is out of date in terms of where customers are at and how they are assessing outcomes. If what I saw at FORA is an indication, then customers are far more measured in their assessments but realize they are only scratching the surface, despite, in some cases, seeing significant gains.

My take

There was a lot to digest on the day and I still have a number of recordings I want to listen to again. The HSBC case is the most comprehensive and detailed example I’ve seen so far of RPA in motion yet with clearly defined objectives in mind that are about realigning customer service to customer needs. In that context, HfS pulled off a masterstroke in making the HSBC story a central plank of the OneOffice discussion. When I consider the number of moving parts that Buckley discusses, it is clear that is way beyond an incremental change program but one which is looking far into the future.

Interestingly, there was no talk in the room of employee reduction but there was plenty of talk about displacement and the opportunities presented by RPA to reskill existing staff, especially those who already have a degree of institutional knowledge.

My abiding impression though was one of genuine hope. Most of those speaking and, a good number in the room, were of an age when they can remember being burned by technology. What they are seeing today in RPA is different. As always though, the results will speak for themselves. In that context, I look forward to hearing more war stories in 2018.

Image credit - HfS Research

    Comments are closed.

    1. Europe and US are a few months ahead APAC in RPA but we’re seeing growth in robotic process automation here in Australia. FORA is a great event and certainly planning to attend next year. Thanks for the insights.

    2. Harald Nehring says:

      I’m really trying hard to not just see screen scrapers in hip “robotic” outfits here. I really do. I see lots of talk using AI for low-level things like identifying entry fields or push buttons on screens, but very little talk to e.g. use it to support front line staff with real deep customer/consumer operational process insight. But isn’t this what drives customer experience so much more than a bit of swivel-chair checking of a bunch of probably heavily disconnected systems? I remember the (broken) promises of late eighties mainframe screen scrapers still too vividly to fall for it that quickly again. What am I missing this time around?

      1. says:

        Well – in which case you might want to ask several of SAP’s customers, one of which presented on exactly the topic you raise, why they were there. Better still, if you had been there you might have learned something useful – as I did. And while we’re talking about broken promises, I’m reminded of what BPR was supposed to deliver in the 1993-1999 time frame by your company. Ring any bells?

        1. Harald Nehring says:

          @Den, my concerns are my own, not my employer’s. You’d be surprised by the amount of people getting excited by RPA over here, actually. Maybe it’s my enterprise architect’s heart that aches at the notion of maybe-not-so-intelligent robots happily hacking away at just the wrong version of the wrong app’s outdated screen all day trying to drive a critical financial/customer service/operational/… process. Now I totally agree that there are useful cases for such technology – but we should be super clear about the types and scopes of processes we want to stitch and rubber-band together that way vs. more long-lasting and resilient methods. It’s the old spaghetti vs lasagne situation again.

    3. says:

      @Harald – look at the technology architectures of the likes of Automation Anywhere, Blue Prism and UIPath, and you will understand that this is BPR from the 80s with technology that actually works. Look at the recording loop technology for processes, the scripting for screen scraping and the central control interfaces they have all established. On top of that, we have interviewed 150 users of RPA in the last 3 months who are experiencing robots at real scale and with really good security protocols. So before you dismiss stuff like this, actually take a deep dive into these emerging products to see that these are far more than light software tools. These are tried and tested with enterprises and consultancies who are doing real hands-on work to stitch together workflows, processes, applications and desktop interfaces to provide a genuine transformation of the digital underbelly for so many enterprises, many of whom have left their underpinning processes festering for years with inefficient manual workarounds and spaghetti code for decades. To be blunt, the technology isn’t the problem anymore, it’s the ability for clients to adapt their culture and way of working to embrace automation…


    4. Posit a smidgen of caution. We are seeing RPA work where BPR had failed because browser manufacturers era has ushered in the open web and text based protocols (such as JSON). Yet don’t be conned into thinking computer security has ‘really good security protocols’. It is still 1980s. Pray the crypto boom lasts for 4-5 years so we can figure out how to add capabilities to tokenize computer security before some evil chaps blow up our powerstation by throwing the generators into reverse or some other horrid calamity stitched together by bots.