Convergence of RPA and AI

RPA is centered around processing high volume task that is repetitive and involves structured information. Users are generating more data than ever and information hidden behind it can be key to success of an organization. Most of the valuable information available today is in the form of semi-structured or non-structured data. It will be difficult for RPA platforms to continue with their success journey for a longer time with their current limitation and challenges.

In last few years, RPA has exceeded the ambitious growth projections made by top market analysts. If RPA is able to get around the problem of handling unstructured information, it will open a new realm of opportunities for itself and will continue to hold the position of the top disrupting automation technology in coming years.

“The world’s most valuable resource is no longer oil, but data”.

And we see why: Data helps create new product, product creates new users, users create more data.

Most of the leading RPA platform providers have already been associating the term ‘Intelligence’ with their product which is as debatable as the definition of the word ‘Intelligence’ itself. RPA architects and developers would be aware that it didn’t live up to the expectation with business problems i.e. working on applications behind Citrix or VDIs, document segmentation, extraction of information dynamic templates and other similar complex tasks. Business problems involving emails, chats, RCA notes by engineers are not even considered for stand-alone RPA without asking business to change the way they collected data. We didn’t even talk about solving image classification, text analytics problem using RPA solution.

RPA providers, to continue with their ambitious growth plan, are looking at union of their RPA platform with Artificial intelligence.

Where will RPA + AI win:

AI will provide greater efficiency to RPA on its traditional turf but given emerging industry trends primary areas (in context of services industry) which will benefit from RPA + AI are:

• Extracting information from unstructured data: The world is full of data, unfortunately the technology platforms that we use today do not understand them like humans do. This is a substantial bottleneck that we face today. According to a study, there is more than 5 hours of video data created on YouTube every second. There is huge amount of data that exists in form of images and scanned document. While AI is still evolving in these areas but it has made a significant advancement to help business start earning benefits. When this data is transformed to structured information using AI, the rest is for RPA to do what it does best. The application area is huge in wide range of domains i.e. Finance, Insurance, eCommerce, Sales, Healthcare, Legal and there is a long list.
• Increase RPA platform efficiency: This is the most interesting benefit RPA will get by befriending AI. AI can help RPA to sense when an application has changed and may be, they can fix themselves some day. AI can help RPA to resolve critical errors and resume without human intervention. AI can also help RPA to allocate resources intelligently for a task.

• Chat & Email: Organization wants to free up their manpower from low skilled transaction without compromising on customer experience. RPA + AI will a be complete automation solution where AI will be used for natural language processing to understand and answer queries just like human does and RPA as an integration layer for interaction with core systems.
• Technical support: RPA and AI will help technical support achieve a new milestone and there is a massive benefit that an organization should expect apart from reduced human intervention and faster response time. A very few to mention are predictive maintenance, self-healing of systems, network security analysis, auto performance optimization, intelligent resource allocation, creation of RCA and knowledge base and many others.

This might sound simpler than it actually is. Artificial intelligence is an evolving field and most of the AI solution that we hear are examples of Narrow AI. They solve very specific problem that they are trained to do. For example, the one trained to play chess can only play chess, Alpha Go, which is Deep Mind’s AI program that won the first ever game against a Go professional with a score of 5–0, can only play ‘Go’ with a constrain that you don’t change the rules 😊

Any product offering to be successful should be able to handle multiple variance of business problems. A small example is of an invoice reading solution that should be able to read infinite varieties of invoices. It is difficult to have a pure & plug a play product that can solve wide range of business problems.

Most of the leading RPA platform providers are already working on popular AI areas such as text analytics, Computer Vision, enhanced OCR and others, one major problem they will face is generalizing AI solution when bundling it with their RPA platform. To overcome this limitation, RPA platform providers are partnering with AI leaders i.e. AWS, Google cloud, IBM to allow business to fine tune and train AI models to adapt to their specific business needs.

This is a promising value proposition but we will have to wait and see how well it will be leveraged to provide real business benefits. It will very interesting to see these two disrupting technologies work together.