How can you optimize the cost and ROI of an automation framework?

Purpose - What is the purpose of automation is import. Ask key questions - Is it going to improve KPI ? What KPI's ? who in organization is impacted ? what factor is the target? - is this Automation focused for certain OKR's (Objectives and Key Results) - Increase automation? straight through processing etc

To achieve optimal ROI, it is important to approach process automation systematically. Begin by comprehensively understanding the entire business process. Initiate the process by eliminating any unnecessary steps, standardizing specific procedures, identifying opportunities for optimization, and exploring automation solutions, such as API integration, to achieve fundamental automation objectives. Robotic Process Automation (RPA) should be considered as a strategic step aimed at enhancing efficiency and reducing manual redundancies.

Automation Frameworks brings the tools and utilities together as a bundle for teams for better efficiency and quality. Teams need to adapt the framework based on the system under test. At times teams take any standard automation framework and try to force fit , which is counter productive. The key to get the ROI for the automation framework is the following aspects- 1. Identify the System under Test 2. Analyse the existing Automation Framework which can be reused . We may need to add different components, as required 3. Consider the whole Test Suite - UT, integration test, API level test, functional tests, common functional tests like performance , reliability etc. 4. Ensure that the Tests are running in the CI/CD build pipeline

Utilize cloud-based testing services to reduce infrastructure costs and improve. Beyond functional testing, consider automating non-functional testing like performance, security, and compatibility testing to further increase ROI. Foster collaboration between development and testing teams to ensure alignment and efficient issue resolution. Regularly assess the ROI of your automation efforts. Calculate time and cost savings, bug detection improvements, and overall testing efficiency. Design your automation framework to accommodate new features and additional tests without requiring extensive rework. CICD,Run tests in parallel to save time which can significantly reduces execution time.

An example of good ROI plan or schema is a good foundation of your Business Operating Model: meaning how you will benefit from adding automation to your existing model. Most likely it has a slow and upward attitude, and if this was communicated properly to higher management/ investors, then you'll find less disappointments to your automation journey.

You need a tool that isn't just for engineers and developers, but that doesn't limit those technical teams. A platform that provides full functionality via APIs so you can deploy automation that scales through code. The tool needs to provide a UI based option as well so your IT and departmental operations teams can drive automation initiatives across your organization. Make sure the tooling you work with is committed to the next phase of process automation: AI-assisted process automation. Tray.io has a great offering that aligns with these items so far - highly recommend checking it out.

Investing in high-priced tools doesn't always guarantee better outcomes. Often, open-source solutions or hybrid models can offer flexibility and cost-effectiveness. It's crucial to assess the long-term implications of your choices; for instance, while a codeless solution might offer quick setup, it might lack the customization a code-based tool provides. Prioritizing compatibility ensures that as your business grows, your tools can integrate seamlessly, preventing potential future costs. Regularly revisiting and updating your toolset can also lead to continuous optimization, ensuring you always get the best value for your investment.

Apart from having proper Open Source Tools, CI, effective design etc, we should focus on cloud based testing platforms like BrowserStack, Sauce Labs can save a lot of time, effort, and money compared to setting up and maintaining your own test infrastructure.

Process Automation- typically OT i am referring here , often it is seen that choice of platform or edge control is either not sufficient due to budget constrain or over specification just due to stereo type thinking or so called TREND. As a orthodox instrumentation engineer of 90's I always prefer balancing of- 1. limiting operations to be automated which are directly adding value to yield. 2. Keep it simple and do not implement to complex controls which requires high maintenance in particular MSME industries of India where OPEX always has constrains. 3. Focus on process parameters which are primary and at basics of entire process performance - mostly in batch process 4. choose the right platform when is comes to choose between DCS and PLC

In the orchestra of automation, the choice of tools and technologies is akin to selecting the finest instruments for our performance. Each instrument carries its own melody, with considerations of licensing fees, scalability, and usability being the notes to harmonize. Weighing the balance between open-source and proprietary, code-based and codeless, or desktop and cloud, is our symphony's arrangement. The selection must resonate with our needs, budget, and skill, ensuring that the automation instruments create harmonious music while optimizing our cost and ROI.

the approach should be iterative improvement; - Continuously assess AI performance with regular updates and feedback loops. - Prioritize data quality, ethical considerations, and security compliance. - Foster a culture of collaboration, learning, and automated testing

In the artistry of automation, adherence to best practices and standards is akin to using a timeless canvas, primed to receive the masterpiece. It begins with a modular, reusable, and data-driven design, each stroke contributing to the reliability and consistency of the final artwork. Abstraction and documentation form the contours, ensuring clarity and comprehension. Error handling and logging act as the fine brush strokes of precision. Coding conventions become the colors that give structure to our creation, while code reviews and refactoring serve as the meticulous touch-ups. Version control and configuration management, the curator's touch, preserve the essence of our masterpiece.

When you have chosen your tool for Process Automation, follow the standards and do not force you to build something different. We had some customers failing to implement Process Automation because they have chosen one tool, but tried to make something different with it. So do not spend too much time to making everything extra and special, follow standards and what the tool provides you, or select a different tool that fits your needs.

This is really important angle that you should be aware of and you should be sure that your stakeholders understand this too. Why is it so important? When doing the right things at the right time you save cost, time, and other things such as reputation. Process automation should be always focusing on the best value you can provide in the shortest period of time. So here it comes to the point where you have to evaluate, how many systems are used, how often the systems gets updated, are these systems easy to automate, because not every system is very automation friendly, also you should be aware of upcoming updates, or plans to change the software used in the company. If you think through all of these points, you'll be the best of the best.

First thing first, not everything is made to be automated. Automation makes sens when metrics and stats have been done in previous phase to measure waste, identify bottlenecks, and point out value added activities from non value added ones. Secondly, you don’t want to fully automate value added activities as you kill jobs in your organisas well as people engagement, therefore create change resistance. Instead, optimize such activities. Thirdly, you want to automate non added value activities, and especially activities that creat bottlenecks in you flows. Remember to measure the ROI when automation is done,… you want to know of automation created value, and by how much. Lastly, use Lean Six Sigma beforehand, during and after delivery.

It is always a good choice to start with a more straightforward, more minor Process to automate. Maybe not the biggest one with the highest benefit would be the best. One step after the other, so you can slightly dive into process automation and see the immediate benefits of it. After that, you can start with a more significant project and steps; sometimes, it is also OK to only automate some things. For example, to automate a process that is used five times a year and costs 5 hours, it is only sometimes valid to spend days or weeks to automate. Process Automation is not always about saving time and increasing security and quality, but choose your Automation Projects with that in mind. Don't do Automation because of Automation.

Automating the right things at the right time can have huge benefits for businesses. But how do you know what to automate, and when? Automating the wrong tasks can lead to delays, frustrations, and even lost revenue. To make sure you're automating the right things, you need to start by assessing your workload. Identify the tasks that are repetitive, and time-consuming, and look for ways to streamline them. Automating these tasks will free up time and resources so you can focus on more strategic initiatives.

In the orchestration of automation, not every note can be a crescendo. Focus must be as precise as a maestro's baton, directing efforts to automate those notes that resonate with the highest value, frequency, and stability. We must eschew the cacophonous temptation to automate low-value, low-frequency, or highly variable tasks. Time and effort must be weighed against the value of automation, considering the melody of manual testing. The expected lifespan and maintenance burden of automation scripts must be harmonized. We automate for the enduring symphony of ROI, not just a fleeting crescendo of a single release or project.

Regular monitoring is key to making sure the automation framework is running as efficiently as possible. This includes monitoring performance metrics like throughput, cycle time, and error rates. It's also important to track the utilization of the machinery and any other equipment associated with the automation framework.

Achieving 100% efficiency from a bot on the first attempt is quite rare unless the automated process is very simple. At @RPATech, we recognize the significance of optimization post-implementation. For several weeks, we closely monitor the transactions processed by the bot and, more importantly, the transactions that require intervention from the "human in the loop". By using the 80:20 rule, we ensure that these unprocessed transactions are added back to the robot queue to improve efficiency.

Our automation framework, a living opus, thrives on continuous refinement. We must be vigilant, tracking metrics like the notes in our score – coverage, execution time, defect rate, and the crescendo of ROI. Feedback from stakeholders, the chorus of developers, testers, managers, and customers, is essential. From this chorus, we draw inspiration to refine our framework, to compose new movements, to correct discordant notes, and to enhance our automation's resonance. The framework is a living, breathing entity, requiring perpetual tuning to maintain its harmony with the ever-evolving landscape.

In the grand amphitheater of automation, we are both audience and performer. To optimize the cost and ROI of our automation framework, we must join a collective symphony. Learning from the seasoned virtuosos, their blogs, books, articles, and webinars, enriches our repertoire. Attending courses and workshops is akin to taking lessons from the masters. Joining communities and forums is our passage into the orchestra. And in return, we take the stage, sharing our own compositions, experiences, and insights. By participating in this vibrant musical exchange, we amplify our knowledge, refine our skills, and contribute to the grand symphony of automation, where each note resonates with wisdom and innovation.

Always have a practice of retrospection at the end of every iteration & or project cycles. Capturing lessons learnt along with the team helps to improvise collectively. As well it gives sense of accountability for every individual, as we are getting to know what went wrong or what could have been done better!

Reach out to your contacts at your sector to share their experiences. They may have faced similar challenges before you. Real life experiences may save you lot of time and avoid wasting your resources.

Automation Frameworks, such as the widely use RE-Framework from Uipath, are great. All the point mentioned are valid, I would like to add one more consideration. You need to plan for how to make everybody adhere to the choosen framework and that can be one of the hardest thing. Make sure you have processes in place to monitor the adherence to the framework. Also, there might be situation was you don't want to use the framework, there should also be a process that defines when excpetions to the framework are permissible.

A part of cost is RPA license. RPA vendors have evolved and there are many models apart from traditional license ( control room + Developer + Runner cost + premium support). License cost optimization effort will depend on what license model your organization is using. other Commercial Licensing models available from vendors are - Fixed cost per year - per function cost ( pay per use) - per minute cost capped at robots - Open source ( pay only if you need support or components like control room )

Don't start by trying to automate a process. Start by documenting and actually running (executing) that process manually, step-by-step. It's only by doing steps manually that you realize which particular steps actually qualify for automation and which are not. Usually - in real-life, we find that only a small percentage of steps in a process are actually worth automating.

One needs to understand major cost components in RPA implementation. Those are as below - Infrastructure - RPA Bot licences - Development Team 1. Infrastructure can be optimised by using 1 VM or machine for multiple processes which executes on different interval as a shared resource. 2. Single RPA bot is capable of running many process. As robots are there to run 24x7. Make sure of maximum utilisation with timely maintenance. 3. Choose processes which are aligned to your business goals, Do proper as is & to be analysis for calculating benefits and improvements, prioritise highest beneficial processes. Keep training citizens developers for performing small automation on their own. These are some ways for cost optimisation.

Set Clear Objectives: Define your automation goals and what you aim to achieve. Choose the Right Tools: Select automation tools that fit your project's needs and budget. Integrate with CI/CD: Automate test execution in your development pipeline. Measure Performance: Monitor test execution times and optimize slow tests. Skill Development: Train your team to use the framework and tools effectively. Regular Review: Assess ROI periodically and make adjustments as needed. Scalability: Make sure your automation framework can adapt as your needs grow. License Costs: Be mindful of the software licenses you need for your automation tools. Opt for solutions that match your needs without unnecessary expenses.