COE Brass Tackscenter-of-excellence robotic-process-automation strategy
I just read another “whitepaper” on why you need an Automation Center of Excellence. Although I agree with the overall sentiment that a COE is required - the sheer number of pages used would indicate that we are creating a government contract to build a wall. I will try to give you all the answers to your COE questions in one simple document.
- Do you need an Automation Center of Excellence to get to Enterprise Scale? Yes
- What department in your organization should ‘own’ the COE?
I jokingly say that it is the department that is willing to pay for it, however the real answer is that the underlying business getting done does not change so the business should own the COE. It is heavily supported by IT.
- How do you pay for a COE? To get started you need to have an executive that has enough budget to fund the startup of a COE. You need to get some wins (possibly losses) under your belt to show you know what you are doing. Over time (6 months) you can transition into a self funded model where the savings that the business is getting can fund the COE.
- How many and what kind of people need to be in the COE?
COE Owner – Determines the automation strategy, approves automation requests and decides on the appropriate “Business Analyst” and “Process Builder”. Background is usually from the business with an eye for innovation and a technical affinity.
Business Analyst – Capable of performing a deep dive on the process and scoping the automation project. Can also be the “Process Builder”.
Process Builder – Performs the design, planning, building and testing of the automation project. Can also be the “Business Analyst”. Background depends on the technical requirements of the automation platform that you choose.
- How much funding do I need to start a COE? If you figure that you need at least two people and funds for licenses and the first couple of projects it is in the range of $400k to get started.
- What is the best tool for us to use for Automation? Catalytic :) - There are many different automation tools available in the market. There are reasons for choosing one tool over the other. Look at total cost of ownership - servers - licenses - maintenance. Look at speed to value - how quickly can you get an automation working for you? Look at Process Automation not just Task Automation. How do you coordinate Bots, People, and AI to get work done? At enterprise scale you will end up with multiple tools for automation, being able to effectively manage them all is a key to success.
- How do we prove that our COE is providing Value? Measure the results. Don’t just measure cost savings, but also look for things like Productivity (freeing people up to do higher value tasks), Risk (fines, compliance issues, cost overruns), and Data (Cleaning up data across multiple systems).
- What should the goal of Automation be in our Business? Save Money? Save Time? To risk sounding cliche - Business Transformation. You want everyone in your organization to be actively looking for ways to reduce effort, and change your customers perception of your company through automation. When your enterprise embraces automation then you will really transform how you do business.
- What should the structure of the COE be in the short, mid and long term? Your COE should evolve from a Centralized COE to a Federated COE to a Democratized COE.
Centralized COE - Great for standardizing the methodologies and framework. Great for getting initial wins and stories to tell. Drawbacks are slower moving than everyone who hears about success wants.
Federated COE - Each major division gets their own COE - different sets of priorities can be accommodated. More scalable than centralized - but less control and governance
Democratized COE - Individuals in the organization have the ability to automate processes. The only way to get to scale that makes a big difference in how the organization operates. Governance and control is greatly reduced and the ability to maintain these automated processes becomes much harder.