Business often suffers from legacy processes that consume mountains of employee time, transferring data from one system to another. It’s often the case that data needs to be moved between modern and legacy systems. Due to the age of these older applications, many don’t offer modern APIs that enable integrations to move data automatically.
This creates a business need to automate the process in another way. Robotic Process Automation, or RPA for short, is a process where a typical human-driven business process is captured by a smart screen recording. This screen recording isn’t like what Camtasia offers, that simply captures the pixels on screen, instead, RPA understands what an input box looks like and source and destination data that make up a workflow.
By being able to automate as many common tasks as possible, human time is freed up for deployment on more valuable things than moving data between systems.
There are a number of existing players in the market, but the big news is Microsoft is throwing their hat in the ring. Recently they renamed Microsoft Flow as Power Automate and while I don’t think the name is as catchy, it does lend itself for the new arrival of RPA.
Most businesses are Microsoft shops and over the last few years, those businesses have transitioned to leverage Office 365. By integrating RPA into Office 365’s suite of services, it means plenty of businesses will look to extract further value from their existing licensing, before looking for a third-party solution.
This week a new Microsoft Mechanics video was released to walk you through the features in action. The example shows the steps involved in using Form Processing (part of Power Automate’s AI Builder), to understand and recognise form elements that arrive via email.
By supplying just 5 example invoices to train the model. The AI learns where the fields are, or could move to. While it understands invoice numbers, business addresses etc, it also understands tables which are incredibly useful.
A pro feature of Power Automate is Predict, this analyses one or more attachments supplied via email and predicts what kind of documents they are (like invoices that match the invoice model).
Invoices can then be workflowed for approval (in the example they use Teams). The workflow continues once an invoice is approved.
Typically this would be the point where the finance team would jump into the legacy finance application to pay the invoice. Instead, using UI Flows, you can enter the invoice details automatically from the data recognised in the emailed Invoice.
UI Flows will automatically drive the interface automatically, as your mouse clicks are recorded and dynamic field inputs are inserted, a lot like a mail merge in word, except this is way more powerful.
While RPA takes a while to implement, it will same thousands of hours per year, needlessly wasted on data input. The other aspect is data accuracy, while humans are great at many things, we do make mistakes, while a trained AI process like this is amazingly repeatable and if you ever change your legacy app, it’s pretty quick to retrain.
While there are paid elements to this solution, I highly recommend you watch the video and understand how powerful this is for any business looking to create efficiencies.
What’s most impressive is that this doesn’t take any programming skills to execute, there’s not a single line of code in this solution, making an in-house build extremely affordable as existing staff could master it.