The Future of RPA: Tapping Into the Power of AI and Data
Author | Min Sun, Chief AI Scientist, Appier
Robotic process automation (RPA) has helped businesses reduce the tedium of mundane tasks for people, giving them the opportunity to work on more rewarding tasks. However, RPA has limitations. The intelligent application of machine learning (ML) and artificial intelligence (AI) can take RPA to the next level.
Software developers have been creating programs to automate tasks for many years, but RPA tools have democratized automation, bringing it into the reach of almost everyone. Drag and drop tools allow frontline workers to take everyday tasks and automate them.
For example, a human resource (HR) officer might receive emails each day for leave requests. They may need to open those emails, extract particular information and compile it into a report that needs to be sent to specific people. It is a repetitive task that is carried out every day and may take a couple of hours. However, an RPA tool can create an automated workflow that scans the email account, extracts the relevant data, compiles it into a report and sends the email faster without hands-on intervention. This frees the HR worker up to focus on higher value activities.
Taking RPA to the Next Step
The capacity for RPA to be applied is significant – most workers have tasks that are the same each day and are simply carried out by following a series of well-defined steps.
However, what happens when the task requires specific human input? It might be a repeating task, but there are enough variations each day, so some human intuition and knowledge is needed. RPA allows variables to be used at each step. For example, it could be used to compile a finance report and the variables could be altered to look at a different account code or financial entity. But what if the list of variables changed every day? What if the report needed to cover the top 10 cryptocurrencies by market capitalization – something that changes daily? Or if that report also needed to alert someone of a budget item that suddenly moved from under-budget to overspent?
It is relatively easy for a person to make a decision and add an extra parameter or variable to a search. AI will enable tools to learn what we want and augment queries and actions themselves. Instead of RPA routines being specific to a task, they can be more generalized and automatically adapt according to the data they are working with.
Many actions, while repetitive, require insight and consideration by a human with knowledge and experience. And this is where the next generation of RPA tools can leverage AI. Humans are very good at answering the question of “What else is important or interesting?”. AI will help RPA tools go further than simply adding more variables to a query. AI will allow RPA to take the next step and answer the question of “What else?”. In effect, the application of AI to RPA will allow these tools to expand the scope of what they can do.
Leveraging Data and AI for RPA
The emergence of GPT-3, Generative Pre-trained Transformer 3, is a powerful technology that uses AI to leverage the vast amount of language data on the internet. By training an incredibly large neural network, GPT-3 is able to understand and generate both human and programming languages with near-human performance. For instance, given a few pairs of legal contracts and plain English documents, it can start to automate the task of writing legal contracts in plain English. This kind of sophisticated automation was unimaginable with classical RPA tools without leveraging data and state-of-the-art AI.
This demonstration of the power of data and AI leads us to a new question. What else can be improved by leveraging data using AI in RPA?
One illustrating example is sending emails in digital marketing. Marketers often send automated emails to customers and prospects. Collating a list of targets and then creating and sending the email is a common use case for RPA, but AI can be used to analyze data and optimize those messages in order to maximize the chance of a message leading to a successful conversion. By applying AI to RPA, marketers can better target their customers with personalized messages, so they are more likely to respond positively.
Next Generation RPA Will Grow From Digital Transformation
This relies on having a sufficient quantity of data and powerful analysis tools that can create models of what to send to whom and the best times. This will see closer relationships emerge between line-of-business managers and IT teams working together to identify opportunities to use RPA.
In the past, humans would analyze this data and form hypotheses about when to send the messages. With AI, you can now conduct the analysis, build the hypothesis, test it and then refine it to ensure the next campaign is more successful.
While RPA has delivered significant benefits when it comes to automation, the next generation of RPA will deliver more benefits through the use of AI and machine learning through optimization. This is not about faster automation but about better automation.
It is likely that as RPA becomes more widespread, businesses will also look at how they can source software bots quickly. Some experts posit that we will see the emergence of “Robotics as a Service” as companies look for ways to deploy RPA through subscription models and so they can access expert assistance.
With businesses undergoing digital transformation, and perhaps accelerating their efforts as they come to grips with the impacts of COVID-19 on their workforces, data is becoming increasingly important. The optimization of RPA will benefit greatly from increased digitization in businesses. As businesses develop data lakes and other new information repositories that are accessible through APIs, it is important to allow RPA tools to have access so they can be optimized.
There are some challenges though. Automation delivers many benefits, but it may not be worth automating every process. Businesses should target the repetitive processes that are most valuable as this is where the most business benefit will be derived.
There also needs to be a plan in place for people. When a task that was previously carried out by a person is automated, a new role or activity needs to be found for that person. While it is true that automation can reduce or remove mundane and unsatisfying tasks, the time that is created through automation needs to be used. Saying someone is now available for higher value work is irrelevant if there is no higher value work for them to do.
When automation is targeted and executed well, the business benefits are quickly realized. The next generation of RPA will go further by not only automating routine tasks but by finding opportunities, through the use of AI, to optimize automation and deliver better outcomes. The application of neural networks will allow RPA to go further and deliver better outcomes for businesses, large and small, as they seek to deliver greater value to their customers.
WE ARE HERE TO HELP
YOU MIGHT ALSO LIKE
Author | Min Sun, Chief AI Scientist, Appier Emotion is one of the most distinguishable human qualities, one that sets us apart from machines. However, it is not out of the realm of possibilities for machines to read emotions and respond accordingly. Increasingly, machines are able to interpret human’s emotional states and adapt their behavior to give appropriate responses – something we call emotional AI, or artificial emotional intelligence (though in the computing field, it is known as affective computing). Here we will explore what it is, how it works, and how it can benefit businesses. Three Types of Emotional AI Emotional AI is the next step in the evolution of artificial intelligence. By interpreting people’s emotions, AI can respond in a much more naturalistic manner, making the interaction much closer to typical human intercourse. There are three main types of emotional AI – natural language text analysis, voice analysis and facial expression analysis. The first two are already quite common, while the third probably attracts the most media attention. Other types of analysis also include mouse movement, eye-gaze, heart rate and electrocardiography, etc. Natural language text analysis It involves AI scanning written text like a review of a product
Cart abandonment remains a major issue in online travel businesses. Proactive customer engagement platforms powered by artificial intelligence (AI), however, are making it possible to catch and convert cart abandoners, effectively and efficiently, through brands’ owned media channels. According to a recent research, 92 percent of all travel bookings in Asia Pacific were abandoned at checkout, with online travel agencies suffering from the highest rate at 95.5 percent, followed by hotels (90.4 percent), airlines (91.6 percent) and vehicle hire companies (88.9 percent). There are many reasons why shoppers abandon cart. Some are just browsing, still comparing prices or simply distracted. Others become frustrated with slow-to-load pages, or are put off by high credit card fees. While some cases of abandonment can be avoided by improving your website or app UX, many can be overcome by remarketing via paid media channels, such as Facebook and Google Ads. While such remarketing can be effective in winning back lost carts and boosting sales, it is also costly, and relies on strong retargeting tech and infrastructure that many companies lack internally. AI, Owned Media and Cart Abandonment Can travel brands tap into the potential of owned media and leverage data to tackle cart abandonment?
The world is awash with data – companies are now able to collect and analyze data from more sources than ever before. However, as a marketer, you would be better off prioritizing quality data over sheer quantity in order to create more effective campaigns. More importantly, while data allows you to find out a great deal about your customers, it also risks breaching their right to privacy. So how do you make sure you are using data to its full potential without compromising the privacy of your customers? The answer is using first-party data. What Is First-Party Data? First-party data is simply data that you have gathered yourself about your customers through both owned online and offline channels. The former includes data relating to customer demographics, behaviors, actions or interests collected from your own websites, apps, CRM, social media channels, and so on. Offline channels, meanwhile, can include customer surveys, feedback forms, call centers and other methods of customer information in your CRM. Why Should Companies Prioritize First-Party Data Now? Namely because you have collected it from your specific audience with their permission, you can vouch for its accuracy. And because you know what you wanted to measure