The unfulfilled promise of automation: DNA matters

The process automation, process mining, and integration categories are blurring. Vendors, by making indistinguishable claims about digital transformation and process automation, suggest similar outcomes. However, what differs are their fundamental approaches, which I call the “automation DNA” that dictates customer outcomes. In many cases, the outcomes are limited to efficiency. But the market demands something more, in the form of end-to-end automation capabilities that push beyond efficiency into other outcomes such as innovation, growth, and true business resiliency.

DNA matters

Every few years, a viral story breaks about how identical twins, separated at birth, discovered one another as adults. These discoveries highlight the power of genetics, which dictate similar outcomes across the siblings’ lives. The documentary film Three Identical Strangers tells the incredible story of triplets unethically separated at birth in the name of science. The tragic tale is also fascinating; Despite three completely different upbringings, patterns of behavior and life outcomes repeated across their lives.

Whether biological or technological, the power of DNA is unavoidable. Today, society is reckoning with how “technological genetics” dictate outcomes, such as how algorithms impact democracy. I have been researching and thinking more deeply about how platform architectures can lead to intended or unintended consequences.

It is time to look closely at the technological genetics of automation platforms. I’ve spent time in leadership roles at the three largest robotic process automation (RPA) vendors and now advise several organizations in the broader automation category. As the process automation, process mining, and integration categories have evolved, the outcomes that vendors promise to have are sounding the same. Despite this, the DNA of automation platforms still leads to different outcomes for customers.

Process automation began with two promises: uniting disparate systems together and unleashing the massive trove of data that sits within your processes and back-office functions. In its current teenage years of growth, unfortunately, both these promises are unfulfilled. 

While task automation tapes systems together, the underlying bot architecture makes it superiorly difficult to address bot fragility, scalability, and most importantly integrating various automation technologies, i.e. process mining, intelligent document processing, AI/ML, bot operational analytics, business value analytics and process lifecycle analytics together.

Equally so, the massive trove of business-specific data is still a distant reality as technology vendors leapfrog to the next best phrase, featuring hyper-automation and an extremely efficient and very flexible autonomous enterprise. 

Digital transformation cannot be achieved through purely tech-enabled, service-led approaches, with the promise to build autonomous enterprises with low code, flexible applications, etc. In the meantime, many customers are simply looking for automation to fulfill the core, initial promises made. 

“Automation is leaving massive business value on the table. In the process
of building digital platforms, organizations can increase efficiency by at
least 200%, and by a further 500% through promoting applications that
increase effectiveness. But the real value bonanza comes from maturing
a platform that gives business flexibility, operational adaptability, strategic
options and resilience at a relatively low cost. This stems from a core focus on the underlying enterprise architecture, which contributes to this value bonanza.”

– Leslie Willcocks | Knowledge Capital Partners | Author & Emeritus Professor, London School of Economics

Better, faster, cheaper

I recently read a post that confidently listed the author’s opinion of what defines automation success: it must be “better, faster, or cheaper” in order to be considered of any value to the business. This is not surprising. “Better, faster, cheaper” perfectly captures the dominant narrative about automation: pure efficiency. The assumption is that companies should create economic value with automation by improving the same tasks that they are already doing.

The efficiency narrative is driven by platform DNA (think enterprise architecture). Its popularity coincides with the rise of RPA and process mining. UiPath, Automation Anywhere, Celonis, and others spend a lot of money (approximately $4 billion) to convince business leaders that there is a gold mine hidden in their processes, and if they only make them better, faster, and cheaper, they will succeed. The approach has value, evidenced by the size of the RPA and process mining markets.

Efficiency is great, but is it enough to help businesses win over the next decade? If history is a guide, the answer is no. In technology, innovation beats efficiency. To that end, business leaders cannot allow the better, faster, cheaper mindset to cloud their view of automation’s potential. In fact, automation offers incredible chances for innovation and growth, pushing organizations into greenfield areas of opportunity.

“Years of working with disconnected systems led to disconnected teams and disconnected processes. Connecting systems with RPA may be trivial, but it has very limited ROI until the business is ready to break the silos and redesign end to end processes. Going after full ROI moves an IA program from a technology play into technology + change management.”

– Maxim Ioffe   |  Global Intelligent Automation Leader, WESCO

A deeper dive into automation genetics

Let’s take a look at a few different process automation solutions in the market today and talk about how their solutions have evolved vs. how their DNA will impact their outcomes:

Robotic Process Automation: Although the last two years have seen multiple M&A rounds with RPA vendors buying up API connectors such as UiPath acquiring Cloud Elements, and Blue Prism nearly merging with Tibco, the underlying architecture of RPA solutions remains bot-based. In other words, the platforms are architected for bots to mimic a series of steps that humans take. RPA platforms, no matter what is bolted on by M&A, are still fundamentally designed to offer better, faster, cheaper because the easiest way to design a process is to copy what humans in your company are already doing, regardless of if the information is accessed by screen scraping or API.

“Automation is more than just about speed and cost. Bot architecture is non-invasive, can help you to get a quick win but may not be a long-term strategic solution, while API integration does not need that level of governance but could have high upfront costs. Business objective determines the choice, in which enterprise architecture is a huge influencer.”

— Ankit Thakkar   | Automation & Finance Digitization Leader, Thermofisher

Process Mining: Aptly named process mining vendors map out what happens in companies and by design take a historical point of view: examine what you are already doing and make it better, faster, and cheaper.

In both of these approaches, use cases that are completely greenfield and innovative are rare. A quick glance at the Gartner list of RPA use cases or process mining use case list makes it clear that each technology is simply rehashing work that people have been doing in companies for decades.

“Business genome mapping is fundamentally a deeper decomposition of business into its elemental workflows.  As is the case with human genome mapping, what benefits might accrue from business genome mapping will be exponential and not just incremental.”

–Manish Garg | Co-founder & CPO/COO, Skan.ai

All you need is love autonomy

There is no doubt that automation powers the autonomous enterprise. But that’s not enough, if one’s definition of automation is RPA. AI-based analytics with predictable and actionable insights, machine learning and behavioral analytics, perdurable governance and possibly more are needed for an effective, truly autonomous enterprise. We are not there yet.

“The significant problems we face cannot be solved by the same level of thinking that created them.”

— Jeanne Ross  | Former Director at CISR (Centre for Information Systems Research),  MIT

MIT Sloan School

This future statement, while achievable – “Ultimately you will have processes that compose themselves on the fly and then decompose themselves”, requires a different level of thinking, architecture, and sound governance. Should we march towards it with unbridled passion, without learning the lessons of how a sub-segment of automation has grown up, we might leave behind a wake of failed customer expectations.

“The need to establish an enterprise architecture blueprint (DNA) is even more crucial than ever before in today’s sprawling Automation landscape. Absence of that leads to many of the automation failures prevalent today. But if your business objective is purely task automation, RPA can do the job. If you are looking for digital transformation, start with the blueprint first.” 

– Rameshwar Balanagu   | Director Digital Strategy & Enterprise Architecture, Avaya

Established companies go through five stages of enterprise architecture maturity — moving from business silos to standardized technology to optimized core to business modularity and then a digital ecosystem.

Searching for innovation DNA

Automation approaches that are built on a foundation of integration appear nimblest in what they can accomplish and most valuable in the long term. You can certainly still accomplish better, faster, cheaper, but you can also do more.

The DNA of the integration approach (you could also call this the API-first approach, as some vendors have) is not a replication of what has been done before, but the ability to weave disparate systems together to achieve new things entirely. While other categories make existing processes better, faster, cheaper, integration platforms beckon companies to don their chef hat and mix up a collection of platforms to see what new recipe they can achieve.

Some have expressed that in terms of “If data is the new oil, APIs are the new pipelines” or still others in terms like Deloitte’s: “APIs are the cornerstone of what is widely seen as the next iteration of business development and revenue generation.” It is no surprise, then, that players in the RPA and process mining categories have been buying up API vendors. Gartner notes that by 2023, “almost all major RPA vendors will offer a broader process automation and integration platform combining screen scraping with APIs.”

Despite this trend, some are correctly observing that the DNA of built-in platforms offer towering advantages over bolt-on. “Mergers and acquisitions often confirm a market need but don’t necessarily solve the customer’s pain point,” observes Workato CEO Vijay Tella. 

In Who Moved My Bots I led a conversation with Microsoft and Hanover Insurance where Prashant Hinge, VP of Automation at Hanover Insurance says, when he makes vendor decisions that need to last for 3-5 years, he is going to choose built-in vs. bolt-on every time.

Today, the market demands automation for end-to-end processes. Last year Gartner noted that they did not recommend RPA as a long-term strategic business strategy. Others have labeled the tech a “quick and dirty” solution. Truth is, RPA and process mining bring efficiencies, but fall short on true business resiliency.

Business resiliency – and success in the next 10 years – will come from the same place it always has: the effective and efficient capture of greenfield opportunities through innovation.

Automation is no exception to this rule.  Enterprise architecture (DNA) matters.