01 May Jargon buster: what does intelligent automation really mean?
Intelligent automation, machine learning, neural networks and AI. The world of work has gone a bit sci-fi recently, hasn’t it?
But while all of these technologies sound great in principle, what do they mean? Often, when you start peeling back the jargon, what remains is a familiar concept with some new window dressing.
One great example is the ‘AI-powered’ chatbot. More often than not, the chatbot interface is multiple choice. When things get complex, a human steps in. Not particularly AI, in either case.
In this post, I want to take a closer look at intelligent automation and see if we can get past the jargon to what lies beneath.
Let’s look at a practical example
Descriptions of intelligent automation and robotic process automation (RPA) can easily become quite jargon-heavy. And there’s no use replacing one piece of jargon with another.
Often, the easiest way to explain something is to show it in action:
- A service centre team has to record specific transactions in a compliance system that is delivered as a browser-based app. Unfortunately, there is no integration between the team’s systems and the compliance application
- Company policy is to record the transaction in the compliance app straight after the transactions are completed in order for the systems to stay in sync. The rep who made the transaction is responsible for this
- There are 22 reps in the team, completing 25 transactions per day, taking 4 minutes each time. Across the entire team, this results in roughly 183 hours of work a week
- Removing this manual task would save the equivalent of 5 full-time employees’ salaries per year
How would intelligent automation solve this problem?
- An hourly report would be scheduled, exporting data from the team’s systems into a generic Excel format
- A programme running on a server works through the Excel file and for each row, logs into the compliance application and enters the data into the compliance system
- A second ‘process integrity’ programme cross-references the compliance data with the team’s system to make sure there are no errors
- A simple dashboard is configured with email alerts ensuring the compliance application has been updated according to fixed parameters
183 hours per week have been handed back to the service centre team. The technical risk is low as the desktop team’s applications are not affected and the project scope is modest.
What are the hallmarks of a good intelligent automation system?
The above is quite a simple example, but it’s a good starting point for understanding how intelligent automation works. In fact, working through a spreadsheet and adding corresponding data records is often used as an RPA training exercise.
If you’re considering applying intelligent automation in your workplace, here are a few things to consider.
Good intelligent automation systems tend to be ‘platform-based’. This means they are delivered as part of an overall solution that can be expanded upon in terms of capabilities and the volume of tasks it can manage.
They will also have excellent systems integration, meaning that these platforms can work alongside and interact with other applications. They will usually be able to support industry-standard interfaces so that data can be exchanged easily now and in the future.
And finally, they incorporate proven automation capabilities, such as:
- Cognitive document automation, where the system can learn from example documents and classify or extract structured data set with little or not intervention
- Robotic processes that are designed to be scalable
- Integrated and pre-configured process intelligence, enabling businesses to continually optimise and refine performance
- Real-time process management and orchestration so teams can review performance – often involving the use of dashboards
Is this really a new concept?
The term ‘intelligent automation’ started being widely used in 2018. It’s not an entirely new concept – workplace processes have been getting automated since the industrial revolution – but the capabilities are far greater in scope now than they were even a few years ago.
How can I get started with intelligent automation?
My advice for people looking to implement intelligent automation for the first time would be to keep it simple. Implement in small steps and manage expectations.
The beauty of the above example is that the only change from the user’s perspective was that a tedious task vanished. Minimal disruption to end-users is always the ideal.
Making large changes to processes typically involves disruption and training – all of which should be avoided if possible. All too often, intelligent automation projects are oversold to users. And when the benefits aren’t realized, people are disappointed.
Begin by clearly defining the problem you’re trying to solve. In the above case, the time-consuming manual entry of data into the compliance application. If there is no problem, there’s no need for automation. Innovation for innovation’s sake is never the answer.
If you have any questions about intelligent automation or want to discuss a business problem you’ve identified, feel free to drop me a line on LinkedIn. And if you want to learn more about the latest workplace solutions, download our report, The Essential Guide to Creating an Optimal Office.