Once they were called the 'death of APIs' and were predicted to entirely replace the API-based integration.
Automation Bots have been the hottest topic of the past decade, the software gold mine, though recently we've seen more and more information about RPA-related problems. Statistics reveal that only 13% of RPA projects advance beyond the pilot stage and almost half of all RPA initiatives fail. Meanwhile, reports indicate the growing importance of API development. An API strategy' goals are not limited to internal operations' optimisation any more but seen as a way to empower businesses to build and participate in digital ecosystems. According to Forrester, API strategy one of 'key foundation for digital transformation'.
'APIs drive agility to optimise customer experiences, create dynamic digital ecosystems, achieve operational excellence, and build platform business models.'
- The Forrester Wave: API Management Solutions, Q3 2020
Business leaders are under severe pressure to accelerate digital changes in response to the crisis. They are searching for technologies to set them on the right course and deliver fast results. Should it be the over-hyped RPA? Or should they select the old-fashion API-led integration? Or do the need both for maximum impact?
It was time to dig deeper into RPA (Robotic Process Automation) and APIs (Application Programming Interface) to see if there is a clear winner. Today, we'll look at each technology's pros and cons, the latest trends, and, most importantly, help identify the best ways of leveraging integration and automation potential.
API and RPA are like non-identical twins - similar and different at the same time. Both technologies aim to reduce or entirely remove human intervention from business processes that deal with information exchange. However, their approaches are entirely different.
RPA is closely related to Business Process Automation (BPA) which emerged in the 90s. While BPA deals with end-to-end complex business processes, RPA robots (or bots) mimic human actions to fulfill specific tasks that require no judgement or insight. Bots carry out routine work performed by people by interacting with applications via their UI (user interface). On the other hand, APIs are used to connect software systems ('digital plumbing') to enable direct communication.
69% of boards report accelerating digital business initiatives in response to the pandemic (The 2021 CIO Agenda, Gartner report). 70% of European executives admitted that they were likely to accelerate their digital transformation now. Companies across all sectors had to swiftly adapt their business models to the new reality. The urgent need to optimise operational efficiency, improve supply chain resiliency, and satisfy customers' digital demands have changed IT's role, making it crucial to business survival and growth. Digital initiatives are not seen anymore as purely IT-related but affecting the overall business and its future. A closer relationship between CIOs and business leaders on the one hand and a tighter budget causing more selective approach towards software tools on the other, lead to IT and business joining forces to achieve maximum benefit from digitalisation.
Technology has always played a critical role in business process optimisation, enhancing performance and efficiency; RPA and API are both popular solutions. Historically, business units preferred RPA as a less technical 'no-coding' option with promised faster results and tech experts' minimal involvement. On the other hand, IT departments, with rare exceptions, favoured API-led integration solutions. How has a closer IT/business collaboration effected those choices? Before we answer this question, let us explore both technologies in more detail.
RPA (Robotic Process Automation) is a common choice for organisations with many routine processes that are both manual and repetitive. Rule-based tasks, data entry and verification can be an ineffective use of man-hours and error-prone. Robots are very well suited to carry out these responsibilities and free up workers to focus on strategic and creative tasks.
As Robotic Process Automation enters its 'Slope of Enlightenment', based on Gartner's 5 phases in the technology hype cycle, let us explore how its role has evolved. The initial over-hype is gradually getting replaced by a far more realistic view of its value, effectiveness and cost; new ways of using RPA in collaboration with other technologies like APIs (Application Programming Interface) are gaining popularity.
The RPA market is predicted to grow from an estimated $1 billion in 2020 to $5 billion by 2024 [Gartner, 2020]. Although Gartner expects that "the majority of 'new' spend will come from large organisations that are purchasing new add-on capacity from their original vendor or partners within the ecosystem", plenty of new buyers are adopting RPA. Large and medium enterprises are attracted by numerous advantages promised by the RPA vendors.
Institute for Robotic Process Automation & Artificial Intelligence (IRPA AI) groups RPA benefits into three major categories [Understanding RPA ROI]:
Without a doubt, a robot is much faster and efficient than a human when it comes to repetitive tasks. With a clear workflow, there are no deviations from the pre-defined process's flow. All automated jobs are completed fast, uninterrupted and without errors. Naturally, accuracy and compliance are among the top benefits of RPA.
According to LexisNexis Risk Solutions, the cost of regulatory compliance reached $181 billion in 2020 for global financial services companies. Thee International Institute of Finance estimates that banks spend 3% to 5% of their total costs on compliance efforts. Automated processes executed by bots are programmed to adhere to compliance standards, policies and regulations. Furthermore, RPA makes regulatory reporting and auditing easier and faster.
Free from the routine tedious tasks, employees can 'focus on higher-value tasks that require more advanced cognitive abilities' [Understanding RPA ROI, IRPA AI]. People could concentrate on what they excel at - think, reason, innovation, managing difficult cases and exceptions, etc. In a sense, this is 'taking the robot out of the person'. [Impact of Robotics, RPA and AI on the insurance industry: challenges and opportunities]
When organisations consider RPA, one of the main drivers is to improve the quality of services offered to their clients. Bots become 'virtual robotic agents' responding to customer enquiries 24/7. Furthermore, by providing these bots with analytical information about customers, organisations can offer a highly personalised experience and achieve 'customer-relationship-systems-on-steroids' [Regtech as a new legal challenge, 2017]. Released human agents could, in turn, be used in areas of customer service where a 'social touch' is precious, for example, life or medical insurance claims. Considering people's emotional state using these services, dealing with human agents would be more appropriate and satisfying. Deploying RPA frees up staff to provide better and more personalised assistance when it's required.
For many organisations, legacy systems are the backbone of their operations. Replacing them would jeopardise not only the smooth running of vital business processes but also might require astronomical investments. RPA does not demand to replace incumbent systems because it automates processes 'as-is'. Still, it's advised to improve and optimise a process before applying RPA; otherwise, it could come back to haunt you later.
Is RPA as good as we've been led to believe? Unfortunately, not. As any technology, especially a new one, there are as many cons as there are pros.
As many as 30-50% of new robotic automation initiatives get stalled, abandoned or even replaced with other technologies (Get Ready for Robots, Ernst & Young Global Limited). Ernst & Young's report shows, that in most cases, the fault is not with the technology, but with the expectations that organisations have. Recent studies demonstrate that ROI from RPA deployment can vary enormously from 30% to 200%. Why do some companies manage to successfully implement RPA and benefit from it, while others fail?
Speed is almost always one of the key factors when adopting a new software platform. Robots can be set-up and deployed much quicker than an API-based network, but does it make RPA a clear choice? Not necessarily! RPA fixes the short-term difficulty but does not help resolve the underlying problem, which is why some refer to RPA as a sticking plaster. According to Ollie O'Donoghue, research VP, IT Services at HFS Research, when non-IT departments start introducing RPA, they're just extending the life of an old legacy system, which means they are, essentially, just buying themselves time. As a cheap and quick fix, RPA often underdelivers, as the maintenance costs are too often not considered from the start.
Historically, IT departments are perceived as a state within a state, an invaluable unit that no one wants to deal with. As a result, when business departments have a need and buying power, they purchase RPA without consulting the IT team. Decisions are taken under a false perception that RPA doesn't need IT, and bots are easy to develop, run and support. Though vendors describe their RPA platforms as no-coding tools, don't be fooled. RPA is just another software solution!
The praised drag-and-drop options are limited. When a complex process is to be automated, programming is unavoidable. RPA solutions need to be installed and, like any other software systems, they need testing, maintenance and support. Too often, when a technology department finally gets involved, the business might discover that the RPA doesn't fit into their overall strategy or simply the team is already overstretched.
It is vital to discuss RPA with all stakeholders at the outset, including both business and IT. It is the only way to make an RPA initiative a success.
'Citizen developer' is one of the latest buzzwords. What are they? According to Gartner, the term is used to describe business users trained to work with no-code or low-code tools. In theory, the advantages are clear - massive savings on a dedicated IT team, fast implemented changes, and increased collaboration between departments. So, is it a real silver bullet?
When it comes to RPA, organisations should be aware that even low-coding is still coding and a certain level of training and technical thinking is required. Secondly, Citizen development is usually a slow and hard cultural change. Thirdly, any software development needs to follow strict testing and deployment processes, not to mention security standards. By giving everyone the power to create and share bots, we might be opening a can of worms that will be extremely hard to manage.
As mentioned earlier, bots free employees from routine manual tasks. What happens to the staff released by RPA? Do these people want to be retrained? Are they keen on changing jobs? Do organisations have a clear plan of utilising the newly available workers? All these questions should be raised and answered from the outset.
It's almost impossible to discuss RPA without mentioning security. Risks can vary from external cyber-security to the internal dealing with sensitive data, unclear roles and responsibilities within the organisation, or lack of security management. Robots are given access to multiple systems, usually with the same hard-coded credentials. Which access should one give to a bot? Can you audit each robot's actions?
In April 2020 Treasury and the U.S. Small Business Administration announced that they would no longer accept Paycheck Protection Program loan applications prepared by bots. 'Without RPAs, the loan processing system will be more reliable, accessible, and equitable for all small businesses'.
If data security is a priority, avoid open-source RPA tools! Instead, select a reputable RPA vendor with well-established security processes and procedures.
Bots' stability is a common concern when it comes to RPA. The truth is, bots are inherently brittle as they mostly rely on a change-prone UI layer. Any system's upgrade that includes interface changes might require re-coding of all the existing bots. RPA robots are not as cost-effective as they may seem. Not only are the implementation costs usually higher than anticipated but also bots require a significant level of support and maintenance throughout their lifecycle.
"You're moving data between existing systems, and if those systems don't ever change, all is good in the world. But that's not really the case. So as systems evolve, things at the back end have to change as well",
- Albert Rees, the head of North America, EPAM.
The above problems make the RPA ROI calculation very complicated. It's the reason why some see RPA as technical debt. The organisation doesn't just pay for it once; there is a constant interest – the effort to manage and maintain it. In real life, some businesses experience that RPA actually slows down innovation instead of accelerating it.
Now that we've listed the pros and cons of robotic automation, let us look at its common alternative - an API-based integration. What is an API? An API is a piece of code that creates an access point to an application. It's an intermediary that accepts requests from the outside, securely passes it to the system and then communicates the response back. For the outside participants of this data exchange, the system remains a black box. APIs enable back-end systems implemented in entirely different ways to communicate and exchange information with each other. The API conception is not new; it was born in the 70s or some even date it back to 1940. What new is the notion of API-led integration as a well-managed, highly secure and well-orchestrated network of API-based connections?
Why do so many companies adopt API-based strategy as an enabler and accelerator of their digital transformation? Most businesses have accumulated many systems and applications over the years, both on-premise and cloud-based. Those systems hold a vast amount of data that, if processed and combined, might become truly invaluable for business's growth. Moreover, the internal processes often rely on information stored in various disconnected systems and, in many cases, require human interaction to copy data from one tool to another, thus creating bottlenecks.
API-led integration streamlines business processes by bridging back-end systems and removing humans from the equation.
APIs were once largely limited to technical domains but have now become a significant engine of business growth.
- from 'What it really takes to capture the value of APIs' 2017, McKinsey Digital.
Let us explore how a business can benefit from an API-connectivity, both from technical and business perspectives.
Salesforce.com generates as much as 50% of its revenue through APIs, Expedia.com - 90%, while eBay- 60%.
- The Strategic Value of APIs, Harvard Business Review, 2015
Are you impressed with the APIs advantages listed above? Is it a perfect solution for any company looking for a transformation strategy? API-led approach is certainly not a silver bullet; it's time to look at the dark side of APIs.
If you search on Google for 'API and RPA', you get almost 7.5 million results. There is a clear pattern, though. Most articles perceive these technologies as competitors, providing a comparison between them. 'RPA vs API: Which is better and why?', 'Pros and cons of RPA platforms vs. APIs', 'Not all automation is the same: The difference between RPA and API' etc.
It's clear that both API-led integration and Robotic Automation have clear benefits and challenges, making it even harder to select the most suitable approach. And why should you limit your choice to one only? Can you get the best of both technologies?
"RPA and APIs are complementary technologies,"
- Herve Coureil, CIO at Schneider Electric.
The initial over-excitement with robotic automation is over. Many companies are entering the 'second wave' of RPA, where they are well aware of the technology, its limitations and advantages and how it can benefit their business.
There is a clear realisation that no single software tool can offer the power, intelligence and capabilities to entirely replace human agents and provide end-to-end automation of complex business processes. Gartner sees the future in 'Hyperautomation', Forrester calls it 'Digital Process Automation', while IDC refers to it as 'Intelligent Process Automation'. Whatever you call the new trend, there is a distinct tendency for consolidating advanced technologies, including RPA, AI (Artificial Intelligence), iBPMS (Intelligent Business Process) to achieve a higher level of digitalisation. Based on our experience, there is one more essential player in the hyperautomation game - the API-led integration.
The border between API-led integration and Robotic Automation is becoming more and more blurred as these technologies are infiltrating into each other's realm.
To overcome the known limitations of traditional RPA, a new variant of robots are entering the market - integration bots. When process optimisation requires automated information exchange between systems with rudimental APIs, these bots use systems UI to access data. We are not talking here only about old mainframes and legacy systems, many service-based software tools consider UI as their primary communication channel and therefore care little about providing API-based access. In this context, integration bots play the role of APIs.
More organisations have realised that the idea of citizen developers building bots is not as simple as it might sound. Firstly, no-code software tools are almost always very limiting in their capabilities. Drag-and-drop can deal with relatively straightforward processes, but for anything more complex, coding is still required. Secondly, turning business users into citizen developers need considerable cultural changes and training. 'Everybody can be a developer', claims Charles Lamanna, the CVP of the Citizen Developer Platform at Microsoft. Sounds too good to be true, at least, yet. And finally, it's one thing to develop robots for your personal use and an entirely different one, for the whole organisation. Numerous risks should be addressed, such as security, data privacy and access control, governance, standards, quality and much more. To resolve the above concerns, some vendors approach robotic automation from a slightly different angle. Their bots are created by professional software developers rather than business users. These professional bots deal with very complex processes and can be developed faster and often significantly cheaper than the traditional ones. Moreover, these robots are more stable and robust, and if anything goes wrong, the support is usually a part of the delivery package.
"CIOs should approach automation deployments holistically, with both APIs or RPA serving an integral role in the process,"
- Steve Shah, vice president of product management and enterprise at Automation Anywhere.
Many API Management platforms providers have also realised the potential of using API-based integration and Robotic Automation together. They partner with RPA vendors to offer connectors for easy and secure API-RPA integration.
Today we've explored the main benefits of Robotic Automation and explained why so many RPA projects still fail. We looked at the importance of APIs, and a key role API-led integration plays in digital transformation success. We also listed the technology's challenges that organisations often forget or neglect. And finally, we examined the new trend towards consolidating API integration and Robotic Automation, previously applied in total isolation.
Though the big question remains - 'which technology should we choose?'. With so much contradictory information available, it is not easy to identify the right solution. Is there a clear winner? No, as it very much depends on numerous factors - the organisation's IT landscape and incumbent ecosystems, available budgets, resources, skills and much more. We propose a different approach. Why not consider a combination of API-led integration and Robotic Process Automation? Knowing their strengths and weaknesses help you strategically apply both technologies and deliver a robust solution to enable your full steam ahead with transformation.
In part 2 of the paper, we'll explore the practicalities of starting API-RPA initiatives.
If you want to make your API and RPA marriage a success, stay tuned.
Why don't you examine how bots can benefit APIs and vice versa?
Select your use case, and implement API-led integration in combination with RPA to automate your process. For experienced MuleSoft users, we'll dig deeper into the platform's advanced capabilities and explore challenging cases!
Let our expert conduct a bespoke API+RPA workshop where your team will spend a day: