Most organisations struggle with fragmentation. Whether that is siloed data, disjointed processes or reliance on legacy systems with poor connectivity. These problems lead to inefficient operations, quality problems and staff wasting valuable time on duplicated work.
Intelligent automation solutions and RPA (Robotic Process Automation) have provided faster, more accessible ways to connect systems. But there are many different options and it can sometimes feel like when it comes down to it, it is too hard to create the integrations needed for your specific processes.
We aren't lying when we tell our customers that we can integrate almost anything. So what integration options are available and how do you choose the best one for your situation?
Intelligent automation integrations typically fall into two categories, either "front-end" or "back-end".
Aka: "RPA", "Screen-Scraping", "Data Scraping", "Web Scraping" or historically "Terminal Emulation".
The RPA revolution has been increasing in momentum and generally relies on front-end methods of connection. This means that the automation software interacts with applications using the same methods that you or I would. For example, if I need to use some customer information, I might open a database, type in a customer reference number, and then click on a button in the user interface to copy the address. The screen scraping software does this task in exactly the same way, short of physically pressing the buttons on your keyboard and mouse.
Most modern tools also have some level of ability to "read what is on the screen". Enabling them to cope with different size screens and minor changes to the information presented.
Can be implemented quickly to do specific tasks
Can be used to extract data from most Windows or web applications regardless of the technology that they are using.
Low code interfaces make setting this up relatively accessible for business users.
Tasks need to be repetitive, predictable, and high volume to be worth automating using this method.
Changes are frequently made to user interfaces when applications are updated, this can easily break integrations created using this method. This reduces reliability and increases maintenance effort and associated costs.
There is a higher processing and programming overhead than other methods.
Without proper controls in place screen-scraping can overload systems unused to "human style actions" performed at "robot speed".
There are sometimes limitations in comparison to what actions a human can do - e.g. some tools are not able to "right click".
Back-end connections usually use an Application Programming Interface (API) to "talk to" applications. Like a Graphical User Interface (GUI), APIs let you interact with the application, but instead of images on a screen, they use words or code to make requests.
APIs simplify the inner workings of an application and provide ways to interact with it that are useful based on the actions that a user might want to do. For example if I wanted to get a customer's address I could use a single line of code (providing a customer number and asking for an address) and the API would return the address without me needing to open the GUI or understand how to find that information in the database. APIs usually have a set of instructions to explain how to use them.
Using an API it is possible to access precise functionality and be sure that you are interacting with the application in a way that was designed and expected. This reduces the scope for errors and increases accuracy.
API connections are more resilient. They won't stop working if the application is updated.
It is more efficient than using a user interface as you are directly interacting with the application and don't need to waste time and energy emulating human actions.
For someone with some programming experience setting up calls to APIs is often faster and easier to implement than setting up front-end connectors.
Modern applications often have prebuilt APIs, but when they are not in place they are not generally a fast or cheap solution. The average cost of developing an API is around £16,000!
Sometimes software vendors charge exorbitant rates for use of the API.
APIs are perceived as harder to set up and are often seen as intimidating by less technical automation creators.
Many automation tools come with prebuilt connectors, these are often the best option if they have the functionality that is needed. However, when a prebuilt connector or existing API is unavailable and front-end connectors are not a good fit, never fear, there are other options!
Sometimes a light-weight bespoke solution is the best choice. The feasibility and cost of a custom solution will depend on the technologies and situation involved, but they are often not the scary endeavour with a big price tag that customers sometimes imagine.
Automation Connector Selector Tool
There are a few other aspects to consider when thinking about automation connection types, so we have created a free Automation Integration Assessment Tool (no registration wall) to help clarify the logic we use when thinking through how best to connect different systems.
Or we are always happy to talk through the options if you'd rather talk to a person.