Automation is increasingly being touted as a potential panacea for digital transformation. But its real value can only be truly realised when it enables business leaders to redesign and digitally enhance work across front, middle and back office; so it’s performed much faster, smarter, more efficiently – and at a meaningful scale too.
And that’s great, but with so many vendors increasingly entering the market offering robotic process automation (RPA) and intelligent automation, comes hype and confusion that masks their ability to deliver what they claim. So when selecting a vendor, you’ll need the foresight to assess their actual technical capabilities. You’ll need clarity because it’s only after proof of concept when you attempt scaling up your intelligent automation programme that any serious tech limitations emerge.
Gartner recognises this issue too, asserting: “Through 2021, 40% of enterprises will have RPA buyer’s remorse due to misaligned, siloed usage and inability to scale.” So to help you better assess vendor capabilities and avoid major regret, here’s four key selection criteria to consider.
RPA revenue to reach nearly $2 billion in 2021 — Gartner
1. Compare and contrast robot capabilities
There are significant differences between the capabilities of RPA and intelligent automation vendors’ software robots. In fact, these can prove the difference between achieving short-term tactical benefits with potentially great effort and risk, or strategic work transformation at enterprise scale with less effort and minimal risk.
On one end of the spectrum are highly advanced software robots trained and run by business users through a collaborative platform, with inbuilt capabilities to safely operate within the full governance of the IT department. These robots are being constantly enhanced to conduct work more like humans. They can intercommunicate to collaborate, swarm to share workloads and operate at an unmatched pace as a digital team – with total integrity and accuracy too.
We’re talking about smart, multi-tasking robots that are increasingly being trusted catalysts at the core of digital work transformation strategies. This is because they effortlessly perform joined up, data-driven work across multiple operating environments of complex, disjointed, difficult to modify legacy systems and manual workflows. And unlike any other robot, they deliver work without interruption, automatically making adjustments according to obstacles – different screens, layouts or fonts, application versions, system settings, permissions, and even language.
These robots also uniquely solve the age old problem of system interoperability by reading and understanding applications’ screens in the same way humans do. They’re re-purposing the human interface as a machine-usable API – crucially without touching underlying system programming logic. This ‘universal connectivity’ also means that all current and future technologies can be used by robots – without the need of APIs, or any form of system integration. No legacy systems are ripped out, and no major process change or mass data migration is required.
This capability breathes new life into any age of technology and enables these robots to be continually augmented with the latest cloud, artificial intelligence, machine learning, and cognitive capabilities that are simply ‘dragged and dropped’ into newly designed work process flows. Ultimately, this means that digital transformation, which would traditionally be cost and resource prohibitive, suddenly becomes achievable. In fact, work that can now be achieved in months, would take IT programs and vast teams of people, years to complete.
Sustainable transformation: why climate change and the digital agenda are more closely linked than ever
At the other end of the spectrum are robots doing keystrokes, running scripts and activities against other pieces of software, which is fine for basic task automations. There are also robots relying on recorded process steps to complete tasks across desktop environments, and that’s fine too. But record-and-deploy robots can’t adjust to any unplanned changes across an ever-changing digital environment, which really limits work performance, productivity and resilience.
For example, a software robot has to navigate a virtual environment like an autonomous car navigates a physical one. Imagine ‘recording’ a journey to work and relying on this recording to navigate the same smooth journey the next day. It would end in a car accident, as the ‘environmental’ conditions would be completely different. Similarly, in the virtual world it would result in a broken robot, an unfulfilled work process and more work for the IT team.
Also, recorded processes aren’t efficient when they run. It’s back to the car analogy. On a recorded trip, a traffic light was on red for two minutes – so the car stopped and waits, but the next day, the light is green but the “recorded” journey says wait for two minutes. Similarly, record-and-deploy robots sit and wait for target systems when they could be proactively working.
2. Gauge overall coding effort
Intelligent automation should enable business users to swiftly respond to market demands, so they don’t want to waste time and effort building robots. It’s far better to swiftly deliver automated work using an intuitive operating system to train and manage robots. We’re talking about simply drawing work process flowcharts that orchestrate an underlying ‘process definition’ language that both robots and humans understand – which also removes the need for coding and any associated risks too.
Any vendor that requires programming expertise to automate each process will actually create code-heavy deployments, extensive debugging effort and high overheads. Also, due to a growing scarcity of coding skills, these automation projects will have to get in the IT queue, which is contrary to intelligent automation’s ‘operational agility’ promise. The IT community’s proper role in intelligent automation is to uphold the necessary governance, security and compliance requirements – and not be burdened with ever more technical debt.
3. Establish scalability and collaboration potential
Intelligent automation-driven transformation at scale is only ever achieved through centralised effort, so insist on collaboration. The best way to scale and compound even more value from intelligent automation is having the freedom to employ robots wherever they’re needed to deliver automated work collaboratively, so it’s shared and multiplied across the businesses. So ask vendors to demonstrate how users can not only centrally design, draw and ‘publish’ new ways of automated working, but share, improve and re-use these automated work assets – anytime, anywhere.
Unfortunately, when any automation technology is distributed across desktops and used in individual contexts – it may help that individual, but it won’t help the whole organisation transform work. This ‘siloed’ approach naturally limits any scaling potential and is obviously not effective in the current climate of constrained and remote workers.
Successful human-machine collaboration needs a collaborative culture
4. Assess security and auditability capabilities
Within any enterprise environment, all intelligent automation activities must be performed most securely, compliantly and transparently, or it becomes shadow IT.
You’ll clearly need a vendor with an operating system that generates a centralised irrefutable audit trail of all process automations, providing granular detail of all robot actions and all training history too. Even better is enabling users to create automated processes, which they publish as a document that ‘is’ the actual process. Change the document and the process is instantly changed, and it’s all securely managed within the central system. This best protects the business from rogue employees and rogue robots.
Desktop automation presents drawbacks because a robot and a human share a login, so no one knows who’s responsible for the process, and this creates a security and audit hole. Another challenge is when a single human user is given autonomy over each process recording this obscures a robot’s transparency and hides process steps. Duplicate this over time and it becomes a potential security threat as there’s almost zero clarity for compliance and governance purposes. Also, any inevitable coding introduces shadow IT – with unaudited process models that represent “back doors”, security flaws and audit failures.
Final thoughts
Ultimately, choosing the wrong intelligent automation technology can limit any digital transformation potential and actually introduce the risk of digital chaos. A better way for intelligent automation is to introduce business-led collaboration, smartly, securely and at scale. In fact, by employing this approach, more than 2000 of the world’s large organisations are achieving major productivity increases via improved ways of working, so they stay agile safe and ahead.
