When we talk about RPA (and we do quite often around here), we're usually talking about Artificial Intelligence (AI), data, and the cloud at the same time. It's likely that in your company, even if you're not yet working on RPA, you're already hard at work on those topics as well. In this article I'm going to look at how these four topics touch and, if done right, reinforce each other.
A well thought out data strategy helps companies to get the most out of their data on the one hand, and to figure out what data they are missing on the other. To fit Robotic Proccess Automation (RPA) into a 'classic' data strategy, you need to do some extra work.
In a classic data strategy, analysts start from a business question. For example, a company is looking for new customer segments or new ways to reach existing segments. Business questions are elaborated in analyses, which result in prediction models that are rolled out to a production environment. To make this possible, data is stored centrally in a data lake. If all goes well, this is where all the data available to the company is stored, in a format that makes it easy to discover new possibilities and connections. Large companies employ their own data scientists to do this. Smaller companies lack this competence, even though they usually do not lack data.
For us, that's a great opportunity to help such a company move forward. Because we have knowledge of RPA as well as data science and we know the interrelationships between these fields well. Even at larger companies, which do have their own 'data people' , we can often add value in the area of data. Firstly, those people are always very busy. And they don't always look at their data with a process focus. We always ask 'How can this data improve the processes?' That's a completely different way of looking at a data lake that is primarily designed for Business Intelligence (BI) and marketing. Companies are sitting on enormous amounts of data, which they often don't look at from a process perspective. When I see that, I immediately start thinking. What can we do with this? What does this data mean for existing processes? What new applications of RPA can we think of based on this data?
You don't need a data strategy and infrastructure to start with RPA. Even without a detailed data strategy, the business case for RPA is good to make. What's more, once RPA is up and running, your software robots will generate data that can in turn be used for analysis, giving you new ideas for improvement. The data generated can also feed back into your data strategy and provide you with insights there. You can go a step further and use RPA in the data process itself. For example, a software robot can search your network for data sources and incorporate them into your data lake. RPA can also take over data in your data infrastructure that is not readable by machines. This often involves older systems that do not support APIs or other integrations. A robot will click through systems all night long, without complaining, to download PDF documents or Excel sheets and then take over in other systems.
I probably don't need to explain the benefits of the cloud. Almost everyone has figured out by now that cloud infrastructure helps you deploy IT applications in a scalable, secure and cost-effective manner. And that for most applications it is much more manageable than in-house servers. Do you want to combine RPA with AI and the large amounts of data that come with it? Then the cloud is your only real option. Storing data in the cloud is completely 'elastic'. That means you can store as much data as you want. You don't have to change your infrastructure if you suddenly want to store 10, 100 or 1,000 times more data. If you pay the bill, of course. But the bill is very reasonable: I've seen prices of 0.01 cents per gigabyte per month. Combining RPA with data and machine learning not only requires a lot of data storage, but also a lot of computing power. That's just not available 'on-premises'. Moreover, the algorithm needs quick access to your data when training a model. The bandwidths that the cloud can offer are absolutely unfeasible on your own hardware.
RPA is designed to make quick work of automating and optimizing processes. If you run RPA in the cloud, you have extra advantages on that front. Identity & Access Management , for example, is easy to set up. In the case of Microsoft Azure, our favourite cloud provider, it also automatically integrates with Office 365. Automatic replication and redundancy are included free of charge. So you can always access your data. Also important with RPA: minimal impact on existing IT. By running in the cloud, with RPA we do not burden the existing IT infrastructure.
There are a lot of cloud providers to choose from. We have chosen Microsoft Azure. On top of the general benefits of the cloud and seamless Office integration, Azure also offers extensive functionality for working with AI and data. Setting this up can often be done without code. Azure thus makes large-scale work with data very accessible. Also for smaller companies. We like to look at things from a process perspective, so also at the power of AI. An RPA robot on Azure, equipped with AI and enough data, is a smarter robot. And a smarter robot helps your processes perform better.
Cloud, data and RPA are therefore domains that cannot be seen separately from each other. If you approach it correctly, they will reinforce each other. It doesn't really matter at what stage your organization is. Are you already far advanced and do you have a data lake, AI in the cloud and your own analysts? Then take a look at how you can use that data and infrastructure for better processes. Not quite there yet? Then RPA can help you make quick steps.