As a CIO, you\u2019ve undoubtedly heard and read enthusiastic discussions around the benefits of automation. Beyond its cost savings due to people replacement, automation offers high value in the form of dramatic improvement to:\n\nProcess efficiency\nCycle time (remember, speed is the new currency)\nProductivity\nQuality (reduced errors)\nScalability\nGovernance and regulatory compliance (requirements can be embedded into automation rules)\u00a0\n\nThe value is easy to understand. But there are four important things to know as you move to automation in order to get it right and achieve the expected value. My colleague, Sarah Burnett, vice president at Everest Group, collaborated with me on this blog and the information you need to know.\u00a0\n1. Know about the automation tools that are available\nAs Sarah points out, there are two \u201chot\u201d categories of automation tools: robotic process automation (RPA) and cognitive computing. It\u2019s important to know their capabilities and constraints.\u00a0\nFor instance, RPA tools automate tasks involving structured data (such as spreadsheets, CSV and XML). They are highly rules-based tools with underlying screen-scraping capabilities. RPA tools typically connect with a computing process through the presentation layer, the user interface (UI). Some of them can connect to other systems through services layer connectors too, but it is the UI interface that is popular due to its ease of deployment. Examples of processes automated using RPA include any kind of form processing, accounts payable, change of circumstance notifications, IT ID deletions and database management.\u00a0\nIn contrast, cognitive tools are intelligent pieces of software with machine learning capabilities. These are based on a variety of machine learning algorithms such as Deep Learning Neural Networks (NN) or Random Forests (RF). Cognitive tools work with unstructured data (such as email, Web content and documents). They learn from experience and expand their knowledge base.\u00a0\nCognitive tools use a variety of techniques to automate processes including tapping into big data, building process patterns, neural ontologies or other semantic techniques. They can build a library of historical patterns or ontologies to describe and understand processes and even predict outcomes based on scenarios that they have come across before. They can infer some operational decisions about the correct way to deal with situations that might not match pre-compiled rules or patterns.\nThey typically connect with a computing process through Application Programming Interfaces (APIs). While intended for unstructured data processing, many also have RPA-style capabilities and can use screen scraping as well to handle structured data. Examples of processes automated with cognitive tools include processing of in-bound documents such as insurance claims, company financials and results processing, legal discovery and IT infrastructure management.\nAs a CIO, it\u2019s also important that you know which of the automation technologies in the market are mature enough to be deployed into production. And, of course, you need to know their price points. \u00a0\u00a0\n2. Know how to deploy automation\nIf you look at automation from a point solution \u2013 such as how to reduce the number of FTEs in a process \u2013 the payback is modest. And deployment can be a long and frustrating path. Instead, rethink your business from a digital perspective. Take an end-to-end view of your organization\u2019s functions and how you can digitize your entire service chain. This gives you a different lens and you\u2019ll have a different framework to understand where and how to deploy automation.\nKeep in mind, too, that the traditional mechanisms organizations use to evaluate technology spend use a functional perspective (infrastructure, network, etc.). But most automation opportunities are end-to-end in nature. Therefore, the traditional evaluation mechanism won\u2019t be able to understand the potential of automation or move to release its potential.\u00a0\n3. Know what\u2019s important in your organization\nIT typically spends a lot of time responding to the squeaky wheels. But it\u2019s important that you know if those wheels are where you should automate a process. Instead of responding to people\u2019s complaints, look instead at volume and where automation can make the biggest productivity change or biggest customer experience change. If you deploy automation in the high-velocity, big payback areas, you\u2019ll likely have savings from the efficiencies that you can use to fund other projects.\u00a0\u00a0\n4. Process design and documentation\nAutomation involves more change than just substituting technology for people. It opens up questions around process design. My advice is to take a more intentional view of the business processes in your organization. Bear in mind that you\u2019ll lose some institutional knowledge as automation replaces people doing the day-to-day rote work. This loss of knowledge or skills can be a hidden cost of automation. It\u2019s important to have good documentation of processes before the people with that knowledge move on. \u00a0\u00a0\u00a0\nTwo more...\nI know I said earlier that there are four things you need to know about automation to get it right. But here are two more factors that are good to understand. First, cloud is both a driver and an enabler of automation. The hybrid environments in many organizations today drive the need for automation to orchestrate services and data integration across the hybrid components of the environment. The cloud is an enabler because many automated solutions reside in the cloud, thus making them easier to deploy.\u00a0\nSecond, keep in mind that you can apply automation tools at different stages of a process. And you can combine them for maximum efficiency. For example, you can combine AI and robotics to automate unstructured inputs end to end.