Originally published in LinkedIn:
The Four Faces of Automation: Tactical, Single Point, RPA, and Cognitive.
Automation is a growing topic all over the world and during the last twelve years that I've been developing software from desktop apps like a dentist program in VB6 to robots that integrate several apps, websites, 3rd party tools, etc. with Automation Anywhere imitating human behaviors, I can admit many things have changed.
Throughout the last months, I have had multiple arguments about what automation is and what not is. Personally, programming is a kind of automation that many people just don't consider because it's becoming a must-skill to have.
Why do I consider programming as a kind of automation? Because each piece of code that is written has the main purpose of automating certain actions of your life, take a look at these examples:
- Waze, you can check the reason for the traffic jam without calling your friend who lives nearby.
- Google Maps, you can discover the best places to visit in Paris without getting some advice from a local.
- Tinder, you can match some beautiful girls or handsome guys on the internet without the extra effort of going to clubs, pubs, and spending a bunch of money on uncomfortable dates.
As you can see, all of them automated certain activities. They involved a lot of human intervention and this is a constant from throwing stones to hunt animals to chat with people across the globe.
Some people would agree or disagree with my viewpoint, but one thing is certain all these apps have machine learning in their cores. Machine Learning allowed them to understand your preferences, needs, and give you the best matches in order to increase their revenues and keep you "addicted".
Nevertheless, let's focus on the main topic of this article, what kinds of automations exist in the market, and how they are different from each other.
There are mainly four kinds of automation: Tactical, Single Point, RPAs, and Cognitive. All of them have their pros and cons, which one fits better for your needs, it's a different topic that you might discover soon.
Automation is not a binary choice between black and white, RPAs and cognitive, single point and RPAs, etc. Furthermore, when you are choosing the best option; you need to be careful because I have heard multiple times that some companies advertise some magical tools capable of automating your tasks without programming skills (false, you must have a basic knowledge of programming to use them properly), other ones would tell you that you can create chatbots to interact with your clients based on your historical data and so on so on; the list of options is growing faster than you can imagine. There is a point when you are confused and you don't know what to choose as CTO, manager, team leader, entrepreneur, consultant, decision-maker, etc.
The previous situation led me to meditate deeper because currently, I need to analyze, what the best solution for my clients could be; do they need an RPA? Do they need a new app? Do they need a macro? Or just need to re-engineer and standardize their processes. After many months of working, with two of these new technologies: RPAs, and a bit of Machine Learning (1), I got some conclusions that I'd like to share with you of these four kinds of automation and how to understand, which one fits your business needs.
- Tactical Automation. This is perhaps, the less recognizable among all kinds of automation since many people don't consider it as automation. It's well-known inside corporations (irrelevant of their size). The most common implementations are the famous Macros. If you have ever worked in a corporation; for sure, one of your friends or colleagues shared with you a magical Excel that can connect to the internet, download data from www.mysap.com, beautify your spreadsheet, and upload it all to SAP in one-shot without major effort. They are excellent when you have a set of actions that are repetitive and connected to Microsoft Office tools like Excel, Outlook, Word, etc. Or 3rd party tools as SAP because all these tools have their own recorders that create codes that you can edit and adapt to your needs. However, they are not exactly the best solutions since they have several limitations, for example, they can barely be optimized, cannot be used by multiple users at the same time (in a shared location like a cloud), are not scalable and integration with other Macros or tools might be complex. Also, let's not forget, these tools are being built in archaic technologies that not even Microsoft wants to support (VBA).
- Single Point Solutions. These are the most common automations all over the world because everyone who has access to a PC or smart device benefits from them. They are custom-made and commonly called apps or websites, for example, WordPress, PayPal, SAP, eBay, etc. I'm sure you can recognize at least any of them. These kinds of services are focused on automating specific situations and most organizations have developed their own apps, CRMs, CMSs, websites, etc. To help their employees be more efficient in their works. These tools tend to be more flexible, scalable, powerful, and capable of higher collaboration even across borders. However, if you need any integration with other services or tools, they might involve a lot of extra-effort since none of them was designed for those purposes (2) and you may need special APIs (if they are available) or sometimes to re-write the entire application (if you don't have access to the source code or use old technology like FoxPro) to reach your goal, which for sure involves a considerable amount of time, resources and money.
- Robotic Process Automation. One of the trendiest topics in all companies and governments. RPAs are technologies for automating repetitive tasks and can create two kinds of robots: Attended and Unattended (I'm going to go deeper in a future article). These tools are known by Efficiency, Consistency, Compliance, and Scalability since they are able to imitate all human behaviors in PCs and follow accepted rules by companies. Also, they can access any program (Word, Excel, Chrome, SAP, etc.), web scraping websites (Facebook, eBay, Amazon, Google, Outlook.com, etc.), connect to VDIs (Citrix), program and run tasks at certain time/day and perform all actions without major changes in the companies' infrastructure or rebuilding internal tools. Further, as Macros, they have their own recorders (desktop, web, Citrix, etc.) for detecting click events, pressing keys, recognizing images or controls, etc. In most cases, you only need a Server and a PC to build, audit, admin, and run them. One important point to highlight, any task to be automated, must be: rule-based, if the process is unstandardized, then re-engineering is the first step to reach successful automation. Now, let's suppose a real-life example to understand better their capabilities: "Daily, you enter your credentials to SAP, download a file from www.acmesystems.com, check if the data provided is correct, upload the data to the SAP box G55; after the process is done, you send an email to the requestor and finally, you update www.supermindset.com/history with the results of your activities." The process described before is a good candidate for automation because all actions are known, there might be some exceptions that can be added to improve efficiency. Also, some tools allow you to go deeper with cognitive options like getting and analyzing data from PDFs or using OCRs.
- Cognitive Automation. This is the last kind of emerging automation where almost any new technology is evolving, for example, machine learning, artificial intelligence, chatbots, among others. These kinds of technologies are capable of understanding human interactions with other beings, predicting your preferences, learning, and improving services by themselves; however, they can be harder and more expensive to implement. You might need a Data Mining Center for the proper data analysis, identifying potential usages and development of algorithms; a luxury for many companies in the short-term. Some companies have opened their clouds and brought specialized services such as Microsoft Azure Machine Learning, Amazon SageMaker or Infosys NIA. These services can empower companies and individuals taking care of the hard work, saving considerable amounts of resources, time, money, and focus. They provide you tools for developing flowcharts, generic machine learning algorithms for analyzing your data, getting better and faster predictions, etc. The previous services are powerful and can help you predict many alternatives, for example, possible investments in Wall Street, tendencies of the generation Z, find an alternative provider for getting sugar if there was a hurricane in the location of your main provider and his production was gone, etc. All predictions are based on your historical data and this is their backbone since they are constantly learning and improving themselves when new data is available.
As you can realize, automation is a wide topic; there are many viewpoints that are constantly changing, but many of us as automators agree, this is a natural evolution from tactical automations (simple macros in Excel) to cognitive automations (predicting the best pair of shoes based on your likes or contacts). Some companies like Google or Microsoft can invest easier in AI services since they have resources and a dedicated workforce, but it's a luxury for the majority, at this point.
Now, I'd like to leave you with a final thought. After, you have a better idea of all these kinds of automations, which one do you think fits better for your organization or personal business? Maybe you need an RPA or just a simple Macro.
Feel free to share your thoughts in the comment section below or ask any questions that you have.
Bibliography.
Bibliography.
Comments
Post a Comment