The Evolution of Automation from Doing Things Right to Doing the Right Things
To begin with, for most organizations, automation is almost always about achieving operational efficiencies. It is about performing an activity or a process in the best possible way with minimum time, effort, and cost yielding maximum output. However, as enterprises make progress in their automation projects, business leaders realize that they must evolve from doing things right to doing the right things.
Businesses want to free their employees from doing mundane and repetitive tasks such as invoice dispatch, payment processing, or credit card reconciliation so that they can be trained to focus on those areas that bring greater value in terms of strategic input or innovative ideas. Thus, enterprises begin their automation journey by improving rule-based processes, through deterministic applications such as monitoring of processes, or workflow orchestration.
As their automation matures to handle the standard enterprise operations, the leaders in the organization begin to look for more opportunities in automation such as how to identify the right vendor at the right time to make a purchase or how to prevent fraud by detecting patterns in transactions. The focus moves to making use of the available data effectively to proactively minimize business disruptions through predictive analytics. With the help of analytics-driven operations such as predictive models and correlational analysis, anomaly monitoring and diagnostic model-based automation, businesses begin to predict failures in advance.
With the evolution of Artificial Intelligence (AI), organizations now look at becoming more smart by adding cognitive capabilities to their automation portfolio. They look at bigger and complex business challenges that automation along with cognitive capabilities could solve such as maintaining zero incidents or reading contracts at speed. Through Machine Learning (ML) and Neuro Linguistic Programming (NLP) and retrospective analysis, an organization could mine processes, build knowledge repositories, and automate actions or provide decision support.
When we look at all the organizations that are traversing the automation journey, the view is hardly monolithic with each of them falling at a different stage of automation. The approach varies based on their position on the automation/AI continuum. Largely, they can be classified as:
Companies that Resist Automation
At the bottom of the automation pyramid are those companies that have been extremely skeptical or wary of considering automation. This skepticism could be due to any number of reasons ranging from uncertainties over the right technology platform that they need to select to a general reluctance due to the complexity of use cases. To remain competitive, these companies will feel the need to embrace automation, sooner or later. In our experience, companies at this stage often seek support in process discovery to understand concepts such as desktop automation etc. rather than a specific automation solution.
At the Deterministic or Predictive stage
At the next stage are the companies that have tried different flavors of automation and are now keen to move towards a more predictive and cognitive automation. They have most likely automated their repetitive and rule-based business processes with Robotic Process Automation (RPA) thereby, achieving significant savings in time and effort. The priority moves from process automation to building a smart environment where AI and automation are implemented with analytics-driven operations to predict failures and create a framework that suggests actions and recommendations.
There are several examples and use cases. Cisco, for example, was looking to transform its customer service experience. While the company tried a Shared Services model, it found that the customer experience was lacking. Cisco runs massive scale of operations with roughly 2 million transactions in a year across 125 services for its customers, just for order management. With AssistEdge, an automation platform from EdgeVerve (an Infosys company), it was able to simplify and automate these processes, thus, reducing order delivery time from four months to eight weeks. The automation enabled cost reductions of up to 80 percent; and reduced two million hours wait time in a year.
Vodafone New Zealand is another example. The company realized that its customer expected a uniform experience irrespective of which Vodafone offering they were dealing with. This was a challenge given that the company had disparate systems for each offering. AssistEdge was able to integrate three separate IT processes into one, enabling a single customer care representative to cater to all three offerings. As a result, Vodafone could completely transform its customer experience processes.
Intelligent Automation Veterans
Today, most organizations fall under the previous two stages mentioned. A very small minority, about five percent of organizations, have truly experienced the power of intelligent automation. With data and predictive analytics capabilities in place, such matured organizations move towards gaining insights that determine the direction of their business. They do this by making use of the cognitive abilities of AI, ML, NLP and pattern analysis to build and manage a repository of knowledge over time. This empowers them to derive evolved patterns and aid business decisions. They not only internalize automation and AI in the way they run their businesses, they also make rapid progress towards larger scale adoption and more complex business use cases.
One great example is our Infosys Nia Contracts Analysis offering. It uses a ML architecture at its core to read contractual documents the way humans would, that is by keeping its context and semantics intact. The system converts natural language into a computable format to maintain semantics and context. There are pre-trained models to help expedite its usage to real-life scenarios. For instance, our customer needs to periodically verify a high volume of contracts (over 25,000) in just three to four days. The process needs to be exhaustive, and there is zero tolerance to inaccuracy. By automating extraction of contractual information, it achieved a massive saving of over 30,000 person-hours a year. Because contract interpretations are standardized, it also helps in early identification of risks.
The Future Ahead
As organizations mature in their automation journey, the workforce will undergo major transformation where employees will be increasingly assisted with AI enabled automation. The future workplace will include both humans and bots and the interface between them could have varied dynamics. Bots could be controlling other bots. Humans could have exclusive command over personalized bots. Bots will progressively become mainstream. The focus therefore, will, shift to developing a strategy around the orchestration and management of the entire digital workforce, in order to ensure that the transformed workplace maintains its functionality, security, scalability and responsiveness.