Companies are on the horns of a dilemma. They signed long-term, managed service contracts for IT or business processes, which took advantage of the savings from labor arbitrage. But now they find that there is significant potential to leverage the new suite of digital technologies that promise improved performance and lower cost. The problem is that that their incumbent service providers often actively resist implementing these technologies, using delaying and obviation tactics, refusing to pass on the savings and/or demanding additional work or other concessions in return for complying. Now that I’ve identified this major issue that many companies face today, let’s look at how they handle this non-alignment situation.
The news of another big round of funding for UiPath, US$225 million series C, and a valuation of US$3 billion created a lot of excitement and amazement in the market. It followed on from Automation Anywhere’s whopping series A funding round of US$250 million in July, which valued the company at US$1.8 billion, and which surpassed UiPath’s earlier series B funding of US$153 million and a valuation of US$1 billion in Q1 2018.
These valuations are phenomenal. In UiPath’s case, the rise from US$1 billion to US$3 billion in less than six months is, I believe, unprecedented. You might think that investors are living on a different planet than us ordinary folks, and that this kind of valuation is plain wrong. I beg to differ.
Investing in the Future of RPA
My case rests on the rapid increase in market adoption and the huge investments that vendors are making in their platforms. As much has already been said about the fast rate of enterprise adoption, there’s no need for me to repeat it again here. Jumping to the second part of my case: RPA today is not the RPA that launched this market three to four years ago. The original developments lacked many of the features that we see today, e.g., computer vision to pick objects on the screen and robust control panels. Similarly, tomorrow’s RPA will be superior to today’s.
As someone who assesses RPA technology on an annual basis, I see a fast rate of product development, not just year on year, but in some cases quarter by quarter.
Everest Group’s “RPA Virtuous Circle” highlights the continuous cycle of developments in the market.
Much has been said of organizations struggling to scale their deployments. I completely agree with this, and for a while I’ve been asking vendors to do something about this issue. I am delighted to see that they have been listening and are investing in features for scaling. These include enhanced robot run time control and management features including intelligent control systems for dynamic workload balancing, auto-scaling, and even identifying processes for further automation. Another major stream of development is turning RPA platforms into the glue that holds together business process management systems (BPMS), different varieties of machine learning, and narrow artificial intelligence. These will ultimately be integrated and will combine seamlessly to provide end-to-end process automation.
While vendors do their bit for scale, organizations should also examine their deployment models for RPA and take a more programmatic approach. Automation is going to be a serious competitive differentiator, and a programmatic approach would significantly speed up organizations’ adoption and realization of desired outcomes. Everest Group’s RPA Pinnacle study highlights some of the approaches that organizations have taken to achieve excellence in RPA.
Of course, these enormous investments in RPA do carry some risks. There is the possibility of tech giants bringing their own RPA solutions to market, in turn pushing out the current RPA vendors. But that wouldn’t be easy to do, as the existing vendors have gained a lot of hard to emulate know how in the past few years. And any one of the existing RPA vendors could be acquired in a major acquisition, but then the investors would get the handsome returns they anticipated…just in a different way.
Taking the Manufacturing Model to Business Processes
Another reason for my optimism about the recent investments in RPA and vendor valuations is that I recently got a glimpse into the future of business automation by looking at manufacturing. On a visit to Siemens Digital, I saw how the concept of digital twin and simulation of manufacturing processes is helping speed up production times and efficiency, even in manual/human processes.
For years, corporate global services functions have attempted to copy manufacturing principles, e.g., adopting Lean and Six Sigma methodologies. Today, they have moved on to automation, which manufacturing adopted decades ago. Having started on automation of global services, enterprises are not going to turn back. They will continue to follow manufacturing’s lead.
Leading organizations are already giving their processes version numbers with supporting documentation, having taken each step through a rigorous Lean Six Sigma methodology. On the automation front, while the focus has been primarily on tactical needs, it will increasingly move to outcomes and the finished “product,” as in manufacturing.
We will see enterprises develop digital twins of their processes or robots, and run complex functions end-to-end in virtual reality before committing to the final model for deployment in the real world. Future versions of RPA will have to support these requirements, and that is where some of the millions of funding will be spent; on product development and advanced features.
Today’s RPA products are paving the way for a far bigger change in automation of global services than we have seen to date. They are the building blocks of the platforms of the future for an inevitable automation journey that every organization will have to take sooner or later. That is why the current group of vendors are so attractive to investors. They are betting not just on today’s growing revenues, but what is to come.
How hot has Summer 2018 been around the globe? Red hot…but not as hot as the RPA marketplace. The speed of evolution in this industry segment is almost without precedent. Firms that had revenues worth tens of millions of U.S. dollars just a couple of years ago are talking about reaching a billion in revenue in just a couple of more years.
So why all the excitement? Some chalk it up to Robotic Process Automation being a clever product idea and others to the even cleverer marketing of sexy robots.
But the reality is that it’s the perfect storm – or heat wave – of innovation and capital intersecting at just the right time.
Of course, it doesn’t hurt that enterprises have already captured most of the potential value from offshore labor arbitrage. But when you combine the need for a new source of cost savings with the acute shortage of labor in the U.S. and Europe, you have a market condition in which enterprises are screaming for automation that allows continued productivity improvements for less money, with less human labor-based effort.
The RPA Virtuous Circle Story
These four keys make up the RPA virtuous circle: More sophisticated software platforms, real value propositions, significant capital infusion, and aggressive buy/build decisions. Let’s unpack each one to get the full story.
More sophisticated software platforms – the software platforms underlying RPA are not new; some of them have been around for many years. But as interest and revenues in the segment grow, the vendors are investing in better software and getting invaluable real-life implementation experience. And great use cases and robust feedback loops will drive enhanced software innovation.
Real value propositions – while a great idea is always fun to talk about, the story quickly fades if the economics are insufficient. In RPA’s case, enterprises are finding real savings and, probably most important, operational improvement. What makes this such an exciting story is that RPA doesn’t apply to just one aspect of the enterprise – it applies anywhere human resources are being deployed for labor-intensive services. So not just G&A functions, but also core business operations.
Significant capital being infused – where there is monetary value creation, Wall Street and Silicon Valley will certainly be found nearby. In the RPA segment, multiple investments in excess of US$100 million have been made. In total, we have seen more than a half billion dollars in investments in just the past six months. These are huge flows of capital, especially considering that in many cases they far exceed current revenues.
Aggressive buy/build decisions – of course, when that much capital is deployed, there’s tremendous pressure to take action to generate real, quantifiable results. The most obvious is to deploy larger sales/account teams to support the growth. But, there will be also significant development needs as use cases expand. We also anticipate that RPA firms will go on a buying spree of niche competitors or companies that increase automation functionality for items like OCR, machine learning, artificial intelligence, and natural language processing.
Right now, the velocity of the Virtuous Circle is increasing…better software, increased enterprise value propositions, and another round of investments.
To learn more about Everest Group’s take on RPA, view the replay of our popular August 8 webinar on the latest developments and implications for enterprises. By registering, you will also receive a a copy of the presentation and deck for download after the webinar.
My recent meetings with the top RPA vendors made it clear that RPA is shifting into new gears of adoption and implementation. But the vendors also made it clear that the true promise of RPA is getting lost in flashy headlines and hype-ridden marketing messages.
Here are my five recommendations for how enterprises can drown out the noise and harness RPA’s real benefits.
Experimentation is Over – The Value is Real
The question “Should I pursue RPA?” has been answered and is widely being replaced with “How can I leverage RPA to gain the most value?” As you see in the graphic below, the benefits of RPA are very real. Nearly every enterprise we have spoken with is seeing real savings – typically around 30 percent lower cost and 30-50 percent improvement in accuracy, cycle time, staff productivity, etc.
Forget about RPA Vendors’ Pitches
Despite all the hype, enterprises must remember that RPA vendors are not selling a digital workforce; they are selling software that can speed up, improve, and support many processes currently performed by staff members. While this is sophisticated software, it’s not a physical entity or an army of robots. It can be tempting to get lost in the imagery, but enterprises need to be careful not to lose sight of what they are getting. Otherwise, they can be left with the feeling that vendors have overpromised and underdelivered.
Ignore the Buzz Words
From OCR to NLP to Intelligent Automation, there’s no shortage of RPA buzzwords. But the labels themselves don’t really matter. What does matter is the ability to identify processes that are using precious staff resources, limiting operational improvement, or diminishing the customer or employee experience. Enterprises should start with the process they want to improve and then approach the vendor with that specific need as the starting point in the context of their overall automation – including and beyond RPA – journey.
Focus on the Operating Fundamentals
The basics of building an enterprise automation capability can seem amazingly easily…until it becomes obvious that it’s not. Some enterprises undoubtedly acquire robots for simple plug-and-play automation. But when mission critical processes come into play, serious and complex issues – like enterprise-grade security and business continuity – come into play and must be carefully and thoughtfully addressed. Don’t allow these issues to become barriers to RPA adoption (as many enterprises do), because, if well implemented, the benefits far outweigh the risks.
Automation Tools are a Must for Business Growth
Automation tools can help enterprises tackle the labor shortage challenge by making their existing teams more productive and retaining key employees by offering opportunities to perform higher-value work. Although cost savings are important, an automation-augmented workforce is key to competing and excelling in the marketplace.
To help you avoid getting caught up in the industry hype around RPA, we’ve created a simple graphic that describes the four key dimensions you should be thinking about. This enterprise automation analysis framework looks beyond vendors’ marketing pitches and addresses questions based on opportunities from your point of view, including:
Business problem complexity – how big and complex is the business process?
Rate of operational improvement – how much of a business process improvement do we want to see?
Solution/technology investment – which of the many different automation solutions should we deploy (considering investment and benefit)?
Operational execution – how do we best implement in your organization?
At the end of the day, however you choose to move forward with RPA technology, start by considering your enterprise’s use cases and business requirements. Then, build the business cases to support them. And then set your automation team loose on an increasingly exciting new set of capabilities.
Insurers are increasingly investing in AI to enhance the customer experience with automated personalized services, faster claims handling, and individual risk-based underwriting processes by empowering agents, brokers, and employees. Our recently released Insurance IT Services Annual Report 2018 found that more than half of insurers are opting to build in-house AI capabilities through hiring, internal training, hackathons, acquisitions, and partnerships with InsurTech companies, while the rest are turning to IT service providers.
Increased InsurTech Investments
The appetite for change within the insurance industry is certainly there. To make that change happen quickly, insurers have been investing in InsurTechs, firms offering technology innovations designed to squeeze out savings and efficiency from the current insurance industry model, to align data and integrate backend systems. Total InsurTech funding reached US$2.3 billion in 2017, a 36 percent increase from the US$1.7 billion recorded in 2016. In 2016, AI and IoT accounted for almost half of the total investment in InsurTech startups globally.
AI InsurTech investment has increased multi-fold since 2016. Seeking access to talent pools, innovative ideas, high speed, and lower cost of innovation, leading insurers have invested in startups including Betterview, Captricity, CognitiveScale, Lemonade, Mnubo, and Uniphore.
And 2018 appears to be spurring even more investments. Indeed, some of the top insurers have created dedicated venture capital arms – e.g., Allianz Corporate Ventures, MetLife Digital Venture, and XL Innovate – to invest in technologies such as voice biometrics, cognitive virtual assistants, speech analytics, telematics, drone imagery, and machine learning.
Research we conducted on 24 leading insurance firms’ investment model suggested that more than 70 percent of their investments in AI InsurTechs are not just from a funding perspective. Rather, they are entering into partnerships with the InsurTechs as a more strategic decision to fulfill their long-term vision of digitalization.
Significant Impact across Insurers’ Value Chain
Process optimization: The majority of the AI InsurTech investments are for automating underwriting policy administration and policy administration, resulting in increased process efficiency. For instance, AXA partnered with TensorFlow to use machine learning to optimize pricing
Product innovation: In addition to fixing processes, insurance companies are partnering with InsurTechs to develop new customized policies and pricing, per user demand through usage-based information. For example, in 2018, Munich Re’s HSB Ventures led a US$16.5 million venture financing for Mnubo, an IoT, data analytics, and AI startup, to build tailored financial solutions to improve the company’s business and facilitate new business models
Customer experience: AI is making traditional claims processing a thing of the past. Companies are pioneering new cognitive solutions that are making the claims process faster, smarter, and more efficient. For instance, in 2018, GENERALI implemented Expert System’s Cogito® technology to focus on registration and claims processing, and to automate the customer email classification, resulting in a swift and smooth claims process and better customer service.
We believe these partnerships create a win-win situation. They give insurers access to the necessary talent pool, latest technology, innovation, and speed they need to thrive, not just survive. And they provide vital to insurers’ ability to compete, and provide InsurTechs with the guidance, infrastructure, funding, and customer base they need to grow.
What are the major areas where companies will focus their spend on technology or third-party services this year? What challenges will impact those investments? In reviewing the trends in 2017, I believe we’ll see more of the same this year and an increase in digital adoption. However, I believe we’re at the beginning stages of a megatrend for the next five years, and I’m calling the start of this phenomenon: I believe 2018 will be the year of IT modernization.
Over the next five years, large enterprises will drive relentlessly to modernize their IT environment. This activity will range from moving workloads out of legacy environments into the cloud, adopting agile and DevOps and investing much more deeply and thoroughly in world-class security.
I differentiate modernization from digital transformation. I see a different set of initiatives occurring often in the same companies, which I characterize as digital transformation. These initiatives often use some of the same technologies; however, they arise from the business and are focused on achieving competitive advantage. The funding, project management, and impact on change management are different in kind and scope. The rise of IT modernization will not slow the need and velocity of digital transformation, which I believe will continue to grow as well.
With respect to digital transformation, we can expect the 2017 trend of digital pilots moving to much bigger programs to continue. However, change management and business model redesign will be a major constraining factor for successful digital transformation, and I believe we’ll see companies start focusing more on managing digital change.
As IT organizations prepare for modernization, they increasingly focus on three main journeys:
The journey to cloud resulting in establishing cloud as the infrastructure of choice
The journey from waterfall to agile
The journey to implement adequate security.
IT modernization will sweep across an organization’s entire IT portfolio, rethinking and restructuring infrastructure, networks, applications, and the process and policies that govern them. I expect IT modernization to drive a profound rethink of the enterprise IT structure as it will both collapse the IT stack and cause organizations to align services by end-to-end functions rather than horizontal functions. In contrast, digital transformation goes end to end and integrates the portfolio. In digital transformation, a company considers pulling workloads and activity out of the enterprise IT function or segmenting it into a different organization that is run end to end.
The results of this modernization will lead to a dramatic decrease in IT costs, while significantly increasing the speed and agility of IT’s ability to react in a timely fashion to business demand. This sudden increase in efficiency will have a dramatic effect on the service provider community, shrinking their existing revenue streams while demanding new skills and capabilities.
The new business models that emerge from this transformation are unlikely, at least at first, to be as profitable as the existing business models based on labor arbitrage. The combination of reduced revenues and lowered margins will place the incumbent service providers in a dilemma with very substantial conflicts of interest. The necessity to protect revenues and keep margins high is likely to make the incumbent service providers poor partners in the emerging digital marketplace.
One potential bright spot for the imcumbents, at least in the short run: although the overall legacy services segment will shrink, I believe IT modernization will result in a set of workloads with new workloads for service providers. For legacy workloads that have not been outsourced and are not ready to be modernized, companies will need to put them into a stable environment. I believe some of those workloads will move to the services market so companies can focus on modernization rather than legacy. This new work for service providers will partially offset some of the runoff that is happening because of IT modernization.
As I look forward to spending trends and challenges for this year, I think Robotic Process Automation (RPA) is hot and will continue to grow in adoption. Artificial Intelligence (AI) is starting to build momentum, and I think it will be red hot in 2018. I see AI being more disruptive than RPA and, therefore, causing greater change management and business model changes than RPA. RPA adoption already was constrained by change management issues in 2017, and I believe AI will be even more constrained by these issues because of its deeply disruptive nature.
We will also see blockchain technology grow in adoption. Although blockchain is truly a disruptive technology, its disruption will focus on specific areas where a distributed ledger can be applied (in comparison to AI, which has a broader set of uses than blockchain). 2018 will see a greater number of blockchain pilots, and some pilots will become programs. However, like AI, RPA and other new technologies, disruptive business model changes will be a major constraint to adoption.
Companies frequently ask us at Everest Group if the benefits a devops team can be delivered in a distributed labor model. In other words, can a company configure a devops team to operate with part of the team in one onshore location and other part of the offshore or in a different onshore location? To be clear, there is currently a significant debate around this question. Many tech companies and new service providers emphatically say devops can’t work in a distributed model. But legacy service providers with large investments in offshore talent factories argue that It absolutely can work and point to examples in which they are utilizing components of a devops model in an offshore and distributed manner.
Legacy service providers have a strong vested interest in maintaining their current offshore factories, which are highly leveraged with cheap junior resources and are working hard to persuade their customers that the offshore models only need an injection of devops technology. However, none of them appear to be running at the productivity level – or even close to the level – of devops teams that are not in a distributed model.
With just seven months to go to the General Data Protection Regulation (GDPR) compliance deadline, many companies still have wholly inadequate data management capabilities. Strict requirements for personal data security, privacy, and the right to erase, among other things, will cause severe headaches for many CIOs not only in the EU but in all regions, as organizations will have to know which data is and is not subject to the regulation, and where in the world it is stored.
No doubt many complex and conflicting scenarios will arise out of GDPR. For example, consider the following data-related issues:
When a request to be forgotten comes in from a customer, how will the organization find all the occurrences of the same data across the vast enterprise IT estate?
Will public and private cloud and other infrastructure providers be able to handle the requirements in a timely manner?
What would be the knock-on effect of a customer asking for his/her data to be erased? What systems will be affected and how would that effect audit trails and other regulatory requirements, such as maintaining company-related data for audit purposes for several years?
These and a multitude of others will take many more years to understand, get guidance on, and resolve. In the meantime, companies must be compliant, or face fines that are the greater of €20 million or 4 percent of global annual turnover.
For those organizations that have not yet prepared for GDPR, the overheads of data management are increasing significantly. For example, they must figure out how to best obtain and maintain personal consent, handle access requests, process revocation of consent and requests to be forgotten, train personnel to know what they can and cannot do with data under GDPR, ensure outsourced services, cloud providers, other suppliers, e.g. in the supply chain, and partners are compliant, and run audits to check the readiness and effectiveness of the provider/supplier/partner ecosystem.
This is where, with its rules-based bots, Robotic Process Automation (RPA) could prove to be God’s gift to the laggards. Scenarios where RPA could be ideal include, but are not limited to:
Running audits of data against consent and revocation databases for compliance
Checking a queue of in-coming consent or revocation requests, and acting upon them, e.g., setting the right flags in systems or actively deleting data while maintaining an audit trail
Producing audit reports
Propagating changes of personal data and related consent across all the systems that hold that data, by cutting and pasting updates and maintaining consent-related databases
The role of AI
As organizations collect more and more GDPR-related data, Artificial Intelligence (AI) solutions could come into their own by helping with risk and impact analysis and reporting:
How many systems will be affected by a GDPR consent and access related change?
What is the knock-on effect on workloads and audits trails? How do these affect other regulatory requirements of data retention?
How many systems will be affected, and what would be the impact on operations and other legal and regulatory requirements?
What is the data security threat level of the day? What is the likelihood of data breaches on a daily/hourly basis, and what preventative measures could be taken?
What security breach has happened and what actions have been taken? Who has been affected by it and must be notified?
Additionally, good governance is an imperative for GDPR. RPA and AI can be used to embed governance in daily operations for enforcing and monitoring purposes.
A new era of data protection is upon us. It is coming at a time when, some would say, that companies have taken far too many liberties with their customers’ data. The full implications for businesses are yet to be understood. But we believe that all organizations that hold or process personal data will experience some disruption in service delivery as a direct result of GDPR. For more on Everest Group’s point of view, please see our latest free publication: “EU GDPR: Is There a Silver Lining to the Disruption?“
Enterprises and their contact center operations are increasingly adopting self-service technologies to improve the customer experience. Over the past several years, chatbots have become one of the most embraced. And it’s easy to see why.
Chatbots are evolving…
First generation chatbots were software programs that could hold a natural conversation on a text-based channel such as chat, email, or social media. Modern day chatbots’ capabilities – interchangeably called virtual agents – have leapfrogged ahead through use of Artificial Intelligence (AI), machine learning, and Natural Language Processing (NLP), which allow them to understand free flowing language and generate real-time responses. They are no longer bound to a single channel, but instead cover the entire ambit of services, including voice support. This helps them fit well with enterprises’ broader omnichannel strategy, as they can deliver a seamless experience across channels.
…and they know your preferences
Unlike earlier versions, today’s chatbots can even contextualize information in a conversation. Here’s an example:
Jim: I would like to know the status of my flight. It’s booking # PNR AZ12356. Chatbot: Hi Jim, checking into it. It appears that your flight’s departure will be delayed by 4 hours, due to scheduling changes. Jim: Can you cancel it and check for an earlier flight in the morning? Chatbot: Confirming…
As per your request, I have cancelled your flight American Airlines 235 from Boston to New York. Based on your past bookings, I have three suggestions for alternative flights on the same day:
Delta 2897: Departs from BOS at 0430 hours and arrives at JFK at 0550 hours
Delta 2471: Departs from BOS at 0600 hours and arrives at JFK at 0720 hours
Delta 3167: Departs from BOS at 0640 hours and arrives at JFK at 0800 hours
Jim: Book one ticket in the second option. Chatbot: Confirming availability. Based on your preferences, would you like me to book an aisle seat, and select a bruschetta omelet with coffee for breakfast? Jim: Yes, that will be great. Chatbot: Excellent. Your flight Delta 2471 from BOS to JFK will depart at 0600 hours on October 30, 2017. Your seat number is 4C. I have sent a copy of the ticket to your personal email id. Have a safe flight.
Notice how the chatbot contextualized the information based on unstructured and more natural language flow, and offered recommendations based on the user’s past preferences. These degrees of evolution have made chatbots much more self-service capable, and are significantly enhancing the experience that contact centers deliver to their client’s customers.
As with all technologies, chatbots come with risks
The end goal for today’s enterprises is to deliver the best possible omnichannel customer experience. Chatbots can help customers solve problems on their preferred channel of communication (voice and non-voice). However, the technology does have shortcomings. The well-known example of Microsoft’s Tay – a Twitter-based intelligent bot that had to be pulled down within 16 hours of deployment due to offensive tweets – highlights one technology gap that needs to be addressed.
Use of Service Delivery Automation (SDA) – which refers to various types of technologies that can automate inputs to a process, the process itself, or the outputs from a process – is surging in the global services industry. When scaling beyond proof of concept, organizations are finding it’s important to bring together the SDA skills and knowledge into an automation Center of Excellence (CoE). Doing so enables the business to develop its SDA capabilities and competencies in a controlled and centralized manner, in turn helping ensure maximum success from the SDA initiative.
Through our research into automation Centers of Excellence, we’ve identified several areas in which organizations struggle.
The right Center of Excellence structure
While there are numerous possible structures for a SDA CoE, we’ve found that a pyramid structure is ideal, as it helps bring the CoE governance in-line with its customers. The pyramid should have three distinct layers, each with its unique set of responsibilities and clearly defined line of communication with the client organization. Clarity around roles and responsibilities across different layers in the pyramid is critical, not only to avoid miscommunications and missteps, but also to help maximize operational efficiency.
The Service Delivery Automation skills demand-supply gap
Demand for SDA skills has far outpaced the talent supply. Some are filling the gap by locating the Center of Excellence in locations with mature, trainable talent. Others are partnering with specialist firms, e.g., technology vendors and service providers, to leverage their domain experience and access to skilled talent, collaborating with startups, and seeking talent from technology groups and professional communities.
Multiple leading global companies are also training their existing employees on SDA. They typically engage technology vendors and/or external consultants to conduct extensive training programs for three to six months. Further, they encourage employees to join and participate in professional networks /communities and other events to learn from other SDA professionals’ experiences. This approach not only helps build internal skills for automation and reduces dependency on hiring from external sources, but also provides FTEs impacted by automation with alternative career paths.
Conventional location strategies don’t work
The traditional offshore-centric sourcing model based on labour arbitrage has limited relevance for SDA. Because of SDA’s unique requirements, organizations are investing in a diversified location portfolio for SDA in order to leverage the best propositions of each. For example, mature talent markets such as India offer a relatively larger talent pool, are suitable for a large-scale centre, and can deliver quick ramp-up pace. Onshore and nearshore locations offer greater depth and breadth of skills, enable greater interaction with business stakeholders, and provide accelerated time-to-market. And co-locating the SDA CoE with existing global services/digital technology centres can help the organization benefit from greater collaboration and economies of scale.