Intelligent Automation: what’s in a name?
Intelligent automation continues to be one of the top technology trends for 2021
In this digital age, efficiency is more important than ever. And even though a lot of businesses digitized their front-office (what the client gets to see), their back-office (what the personnel gets to see and do) is often still very archaic.
As a customer, you can easily fill out an online form to request a consumer loan or make an insurance claim. Yet, it often still takes several days before your request is actually handled because even basic tasks, such as checking if all the information is there, are still done manually.
But let’s face it, this is no longer something customers accept from their bank, insurer or accountant in 2020.
That is why the rise of intelligent automation (or hyperautomation) is for the second year in a row in Gartner’s list of “Top Strategic Technology Trends for 2021”.
But what is ‘Intelligent Automation’ and how can it help your business?
What is intelligent automation? A game-changer.
The world is already familiar with automating some simple repetitive tasks thanks to classic rule-based automation such as RPA tools. This is done by hard-coding specific rules in a computer program.
Therefore, such a program is good at handling the same structured information over and over again.
Intelligent automation helps to automate more than just repetitive tasks, it also helps to automate thinking (or cognitive) tasks. With Artificial Intelligence at the core of Intelligent Automation technologies, there’s no more need to hard-code specific rules to follow.
You just need to feed the machine information and it automatically learns from feedback and new examples. This is why Intelligent Automation is a game-changer. Thanks to this technology, a machine can also apply what it has learned on unstructured information.
|Classic Automation||Intelligent Automation|
|Repetitive tasks||Thinking/cognitive tasks|
|Structured information||Unstructured information|
|Hard-coded rules||Learn from examples|
As the vast majority of data is unstructured, you can easily imagine that this is a game-changer with a huge potential to automate basic administrative tasks, instantly and 24/7.
Intelligent automation is a stepping stone towards General Artificial Intelligence
Artificial intelligence has left the academic laboratories and can be found in several applications we use every day, for example, the facial recognition software to unlock your smartphone.
Or the software that moves some of the emails that you receive directly to your spam folder. These, and many other examples that we know today, are referred to as ‘narrow’ artificial intelligence. The individual applications have learned to be very good at doing one specific task.
On the other side of the spectrum, we have ‘general’ artificial intelligence. Visionaries and dreamers explain general artificial intelligence as the ability of a machine or computer program to think like a human.
Simply put: one machine will be able to do anything a human can do. According to famous futurologist Ray Kurzweil we are still two to three decades away from reaching general human-level artificial intelligence or ‘Singularity’.
Just like the stepping stones to cross a river, there will be stepping stones to go from narrow to general artificial intelligence. Intelligent automation (aka hyper-automation, intelligent process automation, or cognitive automation) is such a stepping stone towards general artificial intelligence.
Intelligent Automation combines all the basic capabilities that any office worker needs. Autonomous driving is a good analogy that also combines various narrow AI-applications that any driver needs, such as road sign identification or finding the optimal route.
The three R’s of intelligent automation
The technological foundations for Intelligent Automation can be summarized in 3 concepts that are easy to understand: read, reason, and rely on.
|Read||Discover (information) by reading it in a written or printed source|
|Reason||Think, understand and form judgements logically|
|Rely on||Depend on with full trust or confidence|
Let’s make this concrete on a typical email from an insurance use-case:
Consists of discovering relevant pieces of information in this email. Data extraction technologies are used to find for example the policy number. Note that the policy number literally appears in the digital text from this email.
If this would have been a paper letter, Optical Character Recognition (OCR) would have been applied first, to transforms a picture of the letter to a digitized text. OCR and data extraction are used in most intelligent automation use-cases.
Means understanding what this message is about and taking a decision on what has to be done. Note that the concepts “car insurance cancellation” or “end of vehicle ownership” do not literally appear in the message.
Technologies such as Natural Language Processing (NLP) as well as Machine Learning classifiers are used to achieve this.
Is simply put an answer to the question “can we trust the machine?” and if this is not the case a person would still need to validate the decision and correct if necessary. Asking a person to validate a decision made by AI is often referred to as ‘bringing a human in the loop’.
Therefore, each decision made is given a confidence score, reflecting the trust in the machine.
Intelligent automation differs from RPA and will power a next wave of automation
Robotic Process Automation (RPA) has been hyped enormously in the past few years giving birth to several unicorns. RPA can be best described as the old excel-macro on steroids, relentlessly executing repetitive macros across not just one but several applications.
A typical use-case is to type information received in a web form into the database of an ERP system. RPA manages to automate many such low value-adding tasks that are based on repetitive tasks, with structured input and fixed rules.
RPA is one of the more advanced forms of classic automation described earlier. Yet, it doesn’t learn from its mistakes and it can’t handle exceptions it hasn’t been programmed to deal with.
The focus of RPA is more on input/output between systems, via screen scraping and mouse/keyboard automation, and less about the actual content itself. Intelligent automation is focused more on the heavy lifting of the actual content.
Typical shortcomings and drawbacks of RPA are:
- No solution for semi-or unstructured content
- No solution if rules are somewhat ambiguous
- Screen scraping and emulation is not future-proof and breaks easily
Intelligent automation overcomes these challenges and will power the next wave of automation.
The benefits of intelligent automation
So why would you opt for an Intelligent Automation solution instead of other (often cheaper) Process Automation solutions?
Intelligent Automation has many benefits, but some typical business benefits are:
- It’s self-learning, the solution continually learns and keeps itself up to date every time it’s provided with feedback.
- It’s easy to set-up, the models underlying the AI technology are already there, the only thing left to do is feeding it training data.
- It has a fast return-on-investment, usually within the year and not requiring CAPEX investment
- It improves overall productivity, by automating boring, repetitive tasks your employees can use that time for more meaningful work.
- It works 24/7, software never sleeps and operates in the cloud, making it accessible from everywhere.
Intelligent automation in practice
Intelligent Automation is here to help you automate cognitive tasks. Just some of the example use cases include:
- Real-time insurance claims handling
- Robotic accounting
- Document collection and validation (ex. for credit applications)
- Digital client onboarding
- Indexation of incoming emails (mailroom automation/email routing)
Contract.fit’s mission is to put Intelligent Automation at everyone’s fingertips.
If you want to learn more about Intelligent Automation and how it could help your business, don’t hesitate to reach out. We’re always happy to share our insights.