Intelligent Automation refers to the fusion of several software technologies, particularly artificial intelligence (AI), operating analytics, and robotic process automation (RPA), to automate business processes that entail both workflows and cognitive understanding.
Businesses are adopting Intelligent Automation at an accelerating speed and expanding scale. In fact, as of early 2019, a survey conducted by Deloitte showed that more than 53 percent of survey respondents had begun transforming their business processes using automation, with the expectation that this percentage would rise to 72 percent by 20211. This rapid growth is primarily due to automation’s ability to deliver meaningful and measurable benefits to businesses in weeks or months, not years, and without disrupting core infrastructure or systems.
Because of the rapid growth in the utilization of automation, we checked in with folks close to the technology to gain their perspectives on the future of Intelligent Automation and what organizations should be doing today. The following are paraphrased responses from the full interviews. The interviewees are a mix of practitioners within their respective organizations’ automation practices, and people in the automation industry itself.
We conducted the interviews in April 2020. While the impacts of COVID-19 were not yet a dominant topic of conversation, you’ll find that the pandemic’s repercussions were top of mind for a few interviewees.
Senior IT Manager
He started and leads Duke’s RPA practice. He is also an ardent Auburn fan.
Head of Consulting—Americas Roboyo
Before joining Roboyo, he worked as a senior leader in digital transformation and was fully accountable for the RPA function, leading the implementation of more than 90 automations. William and his wife of almost 20 years enjoy hiking, traveling, and spending time with their three children.
Director, Automation Consulting
With an extensive consulting background, Eric started with large-scale enterprise resource planning (ERP) implementations before moving to analytics and now automation. Eric actually auditioned for “American Idol.” Unfortunately it didn’t launch him into a singing career, which the Intelligent Automation industry is all the better for.
Director of Sales
He’s a technologist with deep experience in collaborating with companies to develop tailored solutions that best fit their needs. Justin and his family enjoy traveling, spending Saturdays cheering on their alma maters, and taking advantage of the local Atlanta foodie scene.
Director of Automation Engineering
American Tire Distributors
Before becoming an automation practitioner, he built up more than 10 years of experience in process engineering and data analytics. Combining a passion for technology, science fiction, and lean angles, he commutes on his electric motorcycle set to warp 5—unless it’s raining.
Manager of Integration Services
He recently became the lead of their RPA practice. Ashwin loves sports. One of his favorite indulgences is watching ’90s cricket games on YouTube.
Agent on Desktop Model
Referring to solutions that continually monitor a user’s actions to identify process improvement opportunities.
Automation Platform/ Integrated Tech Set
Pre-integrated cognitive software intended to increase the extent to which processes and business functions may be automated.
Center of Excellence (CoE)
A team of skilled knowledge workers whose mission is to provide their organization with best practices around a particular area of interest.2
Allows developers to create code without having to know a programming language via intuitive interfaces.
A specific subset of AI in which a rule-based decision engine uses algorithms to review historical data to revise its decision logic.4
The act of analyzing business processes to find ones that may be candidates for automation.
What technical advancements in the Intelligent Automation space excite you most? What are the key benefits you anticipate?
We see RPA as a point-in-time solution, or a stopgap on our path to more cloud-based services. Things that are current uses of RPA will be supplanted by inherent functionality in final solutions. For instance, our ERP will take on the activities we are currently supplementing RPA with. That doesn’t mean RPA ever goes away. It’s just never the end state.
These are the benefits of our “native tool” approach:
- By investing in the native tool, we invest in standard APIs and tools rather than less predictable bots.
- The scalability of bots is limiting. Meaning, if we need another bot, we need another virtual desktop, etc.
Due to technology and advancements, the total cost of ownership is becoming more reasonable, especially if you are looking at an entire platform. That helps drive down the “per application” costs.
Other ways to reduce cost include utilizing solutions that provide prebuilt development of specific-use cases along with native or out-of-the-box analytics that can monitor and showcase the “health” and efficiency of the automation for the business. Also, some vendors specialize in certain types of automation along with intelligent document processing (IDP)—leading to faster value realization, depending on the use case, and more than just basic RPA.
A key benefit will be the “people side” of technology advancement: I see an increase in openness and understanding of companies’ culture to use these technologies more and more. What will be interesting to watch (and has already started to some extent) is workforce talent changing to match the change in capability. It’s becoming more about the management of processes than knowledge of process execution.
In general, I like to compare what is happening now, with the evolving methodologies and automation of digital work, to what manufacturing went through in the mid-20th century as lean process improvement and the automation of physical work began to refine the mechanizations of the industrial revolution.
I’m really excited about the notion of autonomous automations. By that, I mean where a process is defined as a series of inputs, outputs, data capture, analyses, and decision points forming a continuous, self-correcting, ideally self-improving feedback loop.
William Falquero and Justin Franks:
Integrated tech sets. The composite solutions allow organizations to realize more full-scale process automation opportunities. For instance, automated code checking that integrates with cloud-based application management systems can accelerate the time it
takes to launch an automation by upward of 20 percent. Process identification integration selection capabilities help companies identify and prioritize which processes to automate next.
Advancements are increasing the scope and speed of what can be automated. Solutions are getting better at automating processes that involve less-structured data. Process identifiers that leverage the “agent on desktop” model are helping us identify more processes without relying on manual and time-consuming process mining activities like manual time studies.
Aside from advancements in the technology is the advancement in the mindset. Many organizations are getting past the “Got it. I know what RPA is.” The executive level is catching on to process automation as a key value driver and, consequently, asking more insightful questions about where and how to leverage their automation capabilities.
I’m excited about the advancements that help expedite the automation journey in different ways. Tools are in place that show the ROI very quickly, helping determine which processes we want to automate. There are tools that can help you manage how you scale your automations.
Additionally, I’m excited about the notion of full-stack automation. By that I mean where there is an input, the automation executes four key process components. It makes a cognitive decision, takes the appropriate action, records data in real time according to data governance standards, and uses logged data to conduct analytics to improve cognitive decisioning next time.
Right now, organizations start with which RPA vendor they like best. Tomorrow, they will start by shopping for the right platform—much like consumers who want to create a “smart home” think about which platform they want to align with and buy the corresponding products.
Automated process mining. Right now, we go off and gather requirements to automate a process. We end up needing to gather more requirements than we thought we would need, sometimes making multiple iterations of gathering requirements. This takes time and increases the barrier to which use cases we can target for automation due to SME availability. I’m very excited about the possibility of a tool that gathers the process automatically.
Cognitive capabilities are really promising. However, it seems that to realize their value, an organization’s automation capabilities must be pretty mature, including in both process and cultural perspective.
What trends in Intelligent Automation are you skeptical of or uncertain will deliver on the hype?
I think most trends are viable, depending on the given organization’s strategy. For instance, the trend of low code opens up the feasibility of creating automations to more than just IT. But opening up creating automation to more than just IT, based on your organizational strategy, it’s not as advantageous. Machine learning and AI are more in their infancy and, similarly, how they fit into an organization’s strategy isn’t as concrete.
Companies are being sold on the idea that they don’t need a data scientist because they can have their business users do it. While you don’t need a team of data scientists, the effective use of these technologies directly by business users is possible only for most organizations through solutions that come with prebuilt or out-of-the-box use cases and/or features. Instead of asking end users to create these advanced use cases from scratch, a solution can provide them the capabilities along with the machine learning that typically would be created by a data scientist. In addition, by partnering with the right resources and/or partners to support the mind shift, an organization can deliver faster on “the hype” by trusting in a team and solution that has done it before.
Higher cognitive tech (such as AI, deep learning, computer vision, etc.). It’s all great, but it still surprises me when industry practitioners point companies to these advanced solutions without first discussing process maturity. Automating bad processes just accelerates failure: more defects, more bad decisions based on more bad data, more waste, more bad customer experiences.
William Falquero and Justin Franks:
The whole approach for process identification needs to change. Backing into what the process should be, based on how it’s being done, impedes our ability to think through how it could be done. We are starting with a process designed initially to be executed manually, and then applying automation to it. We need to change our paradigm so that the starting point of the process isn’t manual execution. For example, cars started becoming much more effective when engineers stopped trying to design them to look just like a horse and buggy.
Currently, the gap between expectation and reality still exists, in that identifying exactly what can be automated for a given process isn’t an exact science. The talent pool hasn’t caught up with demand, making skilled resources hard to find. Additionally, the tools themselves are changing rapidly, making it difficult for the talent to stay on top of the field. This rapidly advancing technology also means that the process put in place six months ago may no longer be the most efficient solution, leading to unnecessary use of bot bandwidth.
Three years from now, what will be the biggest difference in how organizations manage their Intelligent Automation capability?
I’d like to see low code implemented at scale. Many platforms have process discovery tools. I’d like to see continued focus on process discovery and code generation versus making the user interface easier.
Measuring value will be placed at the forefront, both for organizations internally and externally. Many organizations still treat Intelligent Automation with an experimental mindset. That will change, and organizations will pick the point (if they haven’t already) where value must be directly tied to the automations they’ve invested in.
Automation CoE will need to evolve from being RPA-focused on tactical automations to platform management, and to account for more end-to-end automations.
The pace and degree to which processes are automated is going to continue to accelerate, and if you don’t automate the automations, you will struggle to keep up. The use of self-correcting or learning automations will be critical to managing the maintenance and support as scale, criticality, and dependency continue to grow. Automating decisions, actions, data, and analytics to create autonomous “intelligent” processes is the key.
Companies will see a mature Intelligent Automation capability as more essential than they currently do. Understanding of the capability will become pervasive across the organization. Right now, a few areas of an organization see the value of RPA, but, generally, it’s not well understood enterprisewide. With more and more use cases, it will become an expectation for each area to have a good understanding of how they think RPA can help.
Today IT has to play goalie to ensure that there aren’t any rogue instances of RPA, which leads to many problems (processes being built wrong, lack of central management). As RPA capabilities mature, there will be fewer rogue use cases.
In the short term, IT should own RPA capabilities because of how technical the tools still are. This may change over time, and you may see more decentralization as the technology and the governance become easier to manage.
The pace and degree to which processes are automated is going to continue to accelerate, and if you don’t automate the automations, you will struggle to keep up.
What key challenges do you foresee arising as Intelligent Automation capabilities expand and mature?
Scalability is and will be a concern. Cloud service migration will help combat this. As the number of automations increases, testing and retesting becomes a challenge. Like in many SDLC efforts, the fact that test environments don’t perfectly mirror production leads to our team anticipating what additional problems an automation might need to overcome and then replicating that issue.
Some organizations go through an automation curve. They start with the high-value use cases, then the less-value but readily identifiable ones. Then they start “turning over stones” and getting back to higher-value use cases, because they start thinking differently about automation capabilities.
As organizations naturally pivot from tactical to strategic-focused automations, they will have to challenge themselves more in asking, “Why are we doing that today?” A more intimate understanding of the nuances of the goals of the given business function will be required. This knowledge will ensure that they are not automating a business process just to automate something—we must ensure that the people and the process supporting the automation are also optimized for improvements. Once an organization can unlock that understanding, of themselves and the tool capability, they are truly able to scale across their organization(s) and provide tangible value in business benefit realization.
Scale. As organizations continue to increase the total number of automations, their ability to manage existing automations while developing new solutions will become increasingly difficult. Having highly skilled resources routinely managing automations (running regression tests, etc.) will become expensive and, ultimately, unfeasible.
William Falquero and Justin Franks:
Value realization is a current problem, and it will only become a larger problem as companies scale their automation capabilities and invest more time and money. Unless they establish a value-realization framework, many organizations will fail to realize the full value of automation because they won’t invest in the right areas and to the right scale.
You don’t realize value at two automations. You start realizing value at 10 (now that you have a workforce). At 20 you bend the curve (assuming you selected processes the right way) of effectiveness. 50 automations require investment, but that’s when your capability starts becoming transformative. You are now automating entire services. Fortunately, there are solutions in place and gaining traction that aid with this. Ones that can be leveraged to help quickly improve ROI.
Prioritizing the obtainable processes. There will continue to be issues around identifying processes that look good on paper to automate. But it’s not until you open the hood on the process, and invest time in process analysis, that you realize it’s just fundamentally not a good process to automate.
Having rock-solid governance in place that accommodates for scaling will be critical.
The next generation of employees will come to market already thinking, “How do I have a machine do this?”
How do you see the perception of the value and utilization of Intelligent Automation changing?
The perception of value will grow as adoption across the organization grows. Federated models have an advantage here in that they can offer less budget and resource bottleneck, and both are the responsibility of the given business unit. Federated models, however, require a strong model and governance structure. But if that’s in place, businesses can run with their own automations. One precursor that allows organizations to get to this maturity is to have well-documented initial wins that demonstrate the benefit of automation and get the business excited.
Given the current global economy, post-COVID-19, organizations want to quantify and realize process improvements more than ever with an Intelligent Automation solution. It’s a pivot to “How can we do more with the same?” or not having to add staff due to an increase in volume or challenges a potential remote workforce can bring to an organization. These events have amplified the value and utilization of Intelligent Automation, and for many of our customers it has accelerated their automation journey due to remote workforce disruptions that may persist beyond the pandemic.
There are inflated expectations of the current capabilities of Intelligent Automations. As we settle into what is actually feasible, the perception will change from “just slap a robot on it” to automation as a fundamental enterprise capability for accelerating value, creating differentiated experiences, achieving productivity goals, increasing resiliency and scalability, and more.
William Falquero and Justin Franks:
RPA capabilities will need to prove themselves as self-funding and integrate into overall business strategy. Process automation needs to be seen as a common capability of the organization, just like product management or sales and marketing. Otherwise, organizations will continue to struggle with business process execution and efficiency, and they will be passed up by those in their industry who have begun thinking about process automation as a key capability, rather than something to simply apply to existing business models. It’s truly a paradigm shift.
Automation needs to be programmatic and integrated with business strategy. The next generation of employees will come to market already thinking, “How do I have a machine do this?” We need to ensure that our governance is robust enough for the scale as well as the new avenues by which processes will be identified.
RPA has the potential to be the catalyst of IT becoming pervasive across the enterprise, not as just the governor of technologies.
What activities would you suggest organizations do now to take advantage of changes to Intelligent Automation as the industry continues to mature?
Get your capability house in order. Some of our existing customers have established their automation capability in a very forward-thinking manner. Because of this, they are able to take a step back and think about the best value of a given automation and prioritize accordingly versus addressing the squeaky wheel, which is still how some organizations reactively identify the next process for automation.
Expand the knowledge base. It’s still important to have a good communication pipeline in place to keep your entire organization informed of what you are capable of providing from an automation perspective, and how that is changing. This leads to a greater cross section of places that automations come from, and the feasibility and potential reusability of them.
Establishing proper governance and the SDLC process, and educating the business users with value drivers, are very important. As Intelligent Automation continues to grow into the cognitive and AI space, the foundation becomes more critical. Ensuring that we establish the roadmap to address scalability, security, and availability challenges is important as RPA demand tends to grow exponentially. Prioritizing and estimating criteria ahead of time are also good first steps. Building frameworks that can be reused in the longer run will also help.
What challenges will organizations face as they travel their Intelligent Automation journey?
Increased investments. To take on more meaningful and strategic use cases takes more time than the simple “swivel chair” use cases. As capabilities continue to push the envelope of what is possible, organization leaders inevitably will see bigger failures that may discourage them. Some are just the growing pains associated with being bold and trying new things. But some of these use cases will need to be treated with extra care (highly strategic or mission critical) when the “fail fast” mentality for mission-critical use cases may not work. The ramifications may be too severe and result in taking two steps back regarding the overall organization’s appetite to automate.
Culture change. People in organizations get nervous about the notion of processes running themselves.
Talent strategies will need to change. As we evolve as a unit of process executors to a unit of process managers, we will need people with skill sets to match.
As automation capabilities mature, do you foresee a change in the types of processes you automate? If so, how?
We aren’t seeing much change in the types of processes that are being automated. We are, however, seeing the evolution of people’s mindset of what can be automated. Once they get a taste of what can be done, business areas start jumping into more and more complex automations.
More than seeing differences in the types of processes we automate, we’re seeing a difference in what Intelligent Automation will mean to our organization. We want to be a next-generation workforce, and we see Intelligent Automation and machine learning as catalysts to propel us in that direction.