Robotics and artificial intelligence (AI) technologies have been with us for some time now but they finally feel as though they have come of age. Organisations are getting to grips with how robotics and AI can help to improve business processes.

However, the whole concept of how humans will work with robotics and AI is misunderstood. There is a mistaken belief that robots are a threat to human jobs rather than a way to deliver higher quality customer service and greater efficiencies across an organisation.

The journey started 20 years ago with intelligent capture when devices acquired the cognition to read text within documents. The key lesson learned from this phase was that you can’t do automation without information: you need to get data into an automated channel before it can be routed to the right person to do something with it.

The next step was to automate the process of actioning the data to achieve an outcome. This had already been proven in the early stages of the AI industry, where, back in the 1950s, a machine was invented that could play chess. In 1997, IBM invented a computer that beat a grandmaster at the game.

While we’re seeing increasing numbers of organisations undertaking pilots and proof of concept exercises with robotic technology, few have achieved anything meaningful with it.

But in the last 20 years nothing this significant has happened. While we’re seeing increasing numbers of organisations undertaking pilots and proof of concept exercises with robotic technology, few have achieved anything meaningful with it.

In fact, we have spotted a real problem relating to CTOs knowing what to do with robotics and how to put them to work to digitally transform their businesses.

Even digital transformation is an industry term people that many people don’t understand. At its simplest level, digital transformation relates to how customers can engage with an enterprise. This doesn’t mean ‘just’ going digital and creating apps – it’s about supporting every channel of communication, whether that’s email, voicemail, or good old-fashioned post.

HMRC and its self-assessment tax service is a great example of digital transformation. It has seen excellent adoption rates and is massively successful. The big challenge with digitising customer channels is the need to prove true and legal identities using e-signatures and trusted IDs via banks, but this can now be done.

A digital mailroom like the one EDM Group built for HMRC captures data from incoming documents and runs a set of business rules that routes it to a group of knowledge workers. It has between 30,000 and 40,000 envelopes coming into its post room daily and the system has achieved a huge reduction in manual effort, taking away the tedious physical work that people don’t want to do.

Organisations can believe that it is too difficult and expensive to use this kind of platform because their business systems are old and hard to use – and still work.

The good news is that those systems may not need to be replaced. The real barrier to automation is how to integrate systems using Robotic Process Automation (RPA). RPA enables organisations to do integration without integrating.

Even though it’s called robotics, RPA does not use a real robot. Instead, it involves a virtual machine. It logs on to the system with a pre-recorded script, it has some intelligence, some auto routing – and types into the system and bridges the gap without any physical intervention.

This solves the classic Catch 22, which is that there’s no point in capturing all of the data within documents into mailroom or emails. The reason you go digital, build an app or put a process online is to have complete digital engagement and interaction with your customer.

However, customers can still print or email that communication and there’s no point in undertaking RPA that can’t automate the whole process – you haven’t solved the problem of what happens when the capture of the data get stunted and stopped.

The assumption that robotics is AI and that you have to tell it exactly what to do is defective – if you try to build a robot-script that covers lots of routing it’s a massive undertaking.

Step forward AI and cognitive intelligence. Semantic meaning and cognition mean you can take a letter from a customer (not just a form) and interpret what type of communication it is. Our technology can discover whether an email or letter is a complaint just by its tone and language, for example.

Of course, knowledge workers are way up the equivalent AI scale. Simulating one second of human brain activity takes 82,944 processors, demonstrating how powerful they are. But knowledge workers don’t always follow the rules. They adapt and learn but they have good and bad days and are difficult to scale.

The objective for the next 5-10 years is, therefore, to use RPA to get people to do their jobs better.Click To Tweet

The objective for the next 5-10 years is, therefore, to use RPA to get people to do their jobs better. This implies a scaling down of manual processes but not a full replacement of human beings. RPA speeds up labour but is not intelligent. Its main advantage is that unlike humans it doesn’t get things wrong.

The next thing to worry about is Superintelligence, a machine that’s smarter than people – but that’s at least 70 years off. Between now and then, RPA is here to assist us humans, not take away our jobs.