Automation must be the most overloaded term in use, and it’s used across the enterprise no matter your role in today’s digital world. Are you a business owner? You want to automate repeatable aspects of the business process to improve productivity and improve agility. A leader? You’re looking for automation to do more with less and reduce the possibility of human error. Technologists look to automation to speed up various operational tasks, and consumers expect new products and services relevant to them to magically (automatically) appear on the device of their choice. The list goes on and on. Thus, the term automation implies different things to different people, depending on their roles and responsibilities in the enterprise. Join us as we pare down this term and align it with the unique characteristics of different enterprise domains.
Before automating anything, there is a fundamental but vital precursor. It is important to automate the right things that are done right. Automating the wrong things, or things that are not being done right, will only proliferate and magnify the problem. For example, “Adversaries R Us” will take great delight in automating the propagation of their latest hack!
So, take a step back to ask some simple but powerful questions:
- Does this process need to be executed in the first place?
- Is this process serving a business function that matters?
- What does the current process deliver from a business point of view?
- Is the process being executed properly?
- Are there ways to improve the process even before automating it?
Automating an efficient process that enables key business functions yielding tangible business value is more meaningful in the long run.
The different domains of the global enterprise that can be automated may be broadly classified as follows:
Process. The term process could apply to the business process executed by various employees in the enterprise, including the back-office processes as well as customer-facing processes. Behind the scenes, this term applies to the IT department, where the engineers, developers and operators are executing processes to enable the development and deployment of new features while ensuring the availability of the infrastructure and platform components and services.
In all these scenarios, process modeling is a critical precursor to automation. A process model provides a visual representation of what is being executed and represents potential failure points and overriding mechanisms. Processes that have been tested and tried with recovery mechanisms in place are prime candidates for automation.
Infrastructure. The term infrastructure could mean any of the components required for the digital workplace, including the facilities, utilities, servers, platforms, end-user devices, networking, storage — in short, everything other than the software applications themselves. The term automation is usually applied to the provisioning and decommissioning of resources, depending on the capacity needed. Traditional mechanisms would mandate the allocation of such resources well in advance, often leading to underutilization. Recently, automation has become about rapidly allocating these resources just-in-time, thus reducing the likelihood of underutilized resources. However, automation in this space also refers to the operational aspects of the infrastructure, such as problem, change, and incident identification, isolation and resolution.
Applications. Applications have a life cycle of their own, starting with their architecture, design, development, testing and deployment, and continuing through life-cycle management. From an applications perspective, automation is about expediting the transition from one phase to another with the ability to roll back as needed. DevOps originated with the automation of the software code being transitioned from the development team to the operations team in a manner that provides full transparency to all the parties involved. Slicing down the applications to more manageable microservices housed in software containers provided the developers the freedom to use resources as and when needed while standardizing what operations needed managing. In the DevOps world, automation has a key role to play in orchestrating the provisioning and decommissioning of these containers, thereby enabling the automated DevOps life-cycle management.
Self-service. Automation of the service provided to a customer actually started several years back with the introduction of voice response units as a supplementary channel to call centers. With the availability of endpoint devices in the consumer’s hands, the concept of self-service has evolved to a whole new level. However, for self-service to work, there are a lot of mechanics needed to ensure the right data is provided with context to the digital consumer.
Interestingly enough, self-service has been humanized with the introduction of digital agents that can have meaningful dialogues with customers. Natural language processing and the continuous learning of new logic paths are integral parts of the automation involved here. However, self-service needs to be based on personas so that the self-service is optimally personalized to an end user. This not only requires automation but also analytics to understand the end user’s preferences and current environment from which the service is being requested.
Supply chain. While the concept of a supply chain may have started centuries back with the barter system, globalization of trade has expanded the ecosystem of trading partners to unprecedented levels, and it continues to grow. The number of parties involved behind the execution of a seemingly simple transaction of delivering a product to a consumer is mind-boggling. Automation of the supply chain to ensure streamlined access to information is vital for key business functions such as order management, inventory control, distribution, logistics, etc.
Controlled access to relevant data across the trading partners is a key catalyst. Automation in the supply chain world involves the rapid and secure exchange of relevant data among enterprises in a timely manner. Data privacy is another key business concern and is driving the need for automating adherence to changes in compliance regulations. The introduction of omnichannels and the need for an integrated back end that provides automated and real-time capabilities are becoming the norm.
Security. There are multiple facets to automation when it comes to security. Reactive automation is about firing up defense mechanisms when a security violation is detected. Preventive automation is about timely measures to ensure no patterns of violation are detected. Predictive automation is about observing various patterns and projecting a security violation that is likely to happen when certain thresholds are reached. Prescriptive automation is about taking remedial measures that prevent the projected violations from happening.
Blockchain has introduced an element of collaborative trust among multiple parties interdependent on one another with no single authoritative source. Blockchain is collaborative security automated; smart contracts are an example.
Additionally, security is no longer viewed as a bolt-on, but rather as an integral part of the solution, service or application. As such, DevOps is now better represented as DevSecOps, where security is embedded and part of the DevOps model outlined above.
Analytics. To the data scientist, automation may simply mean artificial intelligence and machine learning. Automating the study of large masses of data to train computers to interpret things as humans do is opening up endless possibilities. The computer is the new superhuman — so human that it is raising issues such as ethics and trust that have been traditionally unique to the human psyche.
And there you have it — automation applied across multiple enterprise domains! With that as a backdrop, it should be easier to assess what automation means to you, and how it affects your role.