State of the Automation Market
“Robotic automation is promising, but it is rapidly reaching the limits of its impact. Companies are increasingly layering in the cognitive layer onto robotic automation, dubbed cognitive process automation”.
- Robotic automation is rapidly spreading across the enterprise processing landscape. According to McKinsey, automation can translate to 20 to 35 % improvement in efficiency, 50 to 60 % reduction in process time, and returns on investment in triple-digit percentages.
- Several advantages of RPA: capacity creation, enhanced customer service, etc., however suffers from a range of limitations: limited flexibility, inability to handle judgement-based tasks, etc. one RPA vendor reports that even its most mature clients automate at most 50 % of back-office processes, and the majority of clients automate far fewer.
- Enterprises are beginning to employ RPA together with cognitive technologies such as speech recognition, natural language processing, and machine learning to automate perceptual and judgment-based tasks once reserved for humans.
- Applying cognitive technologies can allow organizations to drive the level of automation to beyond 50%; Analyst firm Forrester suggests that a best practice for installing RPA is to “design the system to potentially link with cognitive platforms”
- Illustrating the benefits of cognitive RPA, a leading global bank used cognitive RPA to automate 57 % of its payments work in the highly regulated area of foreign trade finance.
Complexity of Managing Cognitive Automation at Scale
“In spite of the promise, companies have been slow to adopt cognitive process automation. Organizations are implementing one off cognitive extensions to their RPA deployments, but are nowhere near cognitizing end to end business processes. Several challenges to ensuring true enterprise scale cognitive process automation.”
- Many large enterprise software vendors have begun to incorporate cognitive technologies into their products, however a complete platform for enabling cognitive process automation at scale is lacking.
- Applications of machine learning that provide greater insight to organizations are proliferating. However, enterprises have been relatively slow to implement cognitive process automation applications. Most companies are in early stage development, using individual pieces of AI, but rarely connecting them into a complete end-to-end automated process, much less into a process flow powered by AI.
- This is because architecting and building custom solutions based on cognitive technologies can be complex, and the required skills scarce.
- Several challenges to realizing benefits from cognitive process automation:
- How do you readily orchestrate various cognitive agents into a seamless end to end process?
- How do you ensure the process is robust, scalable and can withstand subversive attacks?
- How do you ensure the cognitive layer provides insights relevant to the decision-making required?
3 Themes – Enterprise Scale Orchestration, Trusted Automation, and Personalization
“Realizing efficiencies from cognitive process automation will require incremental effort from organizations in areas that are new to plain robotic automation: orchestration, enabling trust, and driving personalization. These are critical ingredients for successful deployment of CPA.”
- Enterprise scale orchestration
- Interoperability of bots, intelligent agents, automation scripts developed in potentially diverse environments
- Orchestration, deployment and administration of the components of a complete end to end process
- Making it easy for business users to rapidly compose, extend or modify such cognitively enabled processes
- Trusted automation
- Management, governance, and language support layer underneath the bots, agents and automation scripts
- Flag or halt a transaction if there are compliance violations or sensitive data is being shared inappropriately between processes
- Robustness, security, ability to detect data attacks – instrumentation to surface potential drifts, bias in recommendations etc.
- End to end business processes cross functional boundaries, and require cross-functional collaboration, thus driving the need for contextual and role specific insights.
- Exception and alert management – effective human-in-the-loop operation can only work if the cognitive agent is provide insights in a personalized manner
- Personalization – Event-based, temporal, and probabilistic (judgement oriented)
- Enterprise scale orchestration