Blog

I think, therefore I’m cognitive RPA

I think, therefore I’m cognitive RPA

Author:Sudhir Sen, Co-Founder of Option3 and Products Head of JiffyRPA

Recently, I was talking to a group of people at one of the industry events in Bengaluru, and I was quite surprised to know that many believe Robotic Process Automation (RPA) is still based on the rules engine and that it needs to come of age. While I do agree that RPA industry is still on the borderline of cutting-edge AI techniques like machine learning, RPA actually hascome of age in the last couple of years with its human-like cognitive capabilities.  

Today, it is not just about automating a few processes. It is about automating even the most complex of processes end-to-endwith unparalleled ease, as well as learning from past actions and displaying self-learning capabilities. So, in our own experience of building an RPA product, we made sure it is far beyond any simple, rule-based automation. JiffyRPA is the closest one can get in terms of ease of installation, integration, intelligence and decision-making capabilities.

Firstly, why cognitive RPA?

If you look at how the market segment is panned out today, you will see that large organizations in industry verticals like Retail, Healthcare, BFSI, Manufacturing, Transportation, IT & Telecom, would need automation like never before, simply going by the data explosion and process complexity that we see today.

#CognitiveRPA #JiffyRPA

 

 

Cognitive RPA and BFSI

 The biggest consumers of RPA today are BFSI applications. This is simply because the industry is highly regulated and there is a dire need for intelligent automation, as fixing the simplest of human or machine errors can turn out to be an expensive affair. This is why cognitive RPA will enhance the accuracy and efficiency of different processes in the BFSI industry especially.

Simple automation of monotonous tasks often eliminates workload and results in enhanced productivity and fuels competitiveness. But that in itself is not enough for large organizations that have complex processes and systems. Banks, for example, where a minor error can lead to a catastrophe. This is where “cognitive RPA” comes into play.

If RPA is a gateway technology, then RPA with cognitive abilities is at the heart of that technology – the robust backbone to digitization. Innovation doesn’t and will not happen by simply integrating RPA with existing IT infrastructure, but what will really add value are bots that can add more decision-making capabilities and augment human productivity.

JiffyRPA: I think, therefore I’m cognitive.

Automation was expected to be the silver bullet to the widespread operational complexities of the business world, but this was never going to be the case unless automation could evolve beyond what it could merely achieve – the basics. Most automation solutions were approached from a short-term solution angle, a somewhat tactical necessity. In doing so, they were failing to leverage the actual improvements to be brought about from a big picture perspective. And in understanding this, our approach towards intelligent automation was born – Automate, Analyze, Accelerate.

Option3 time and again proved to our customers how limiting it would be to view RPAas a tactical tool, when it could be a real strategic asset. JiffyRPA’s cognitive capabilities, for instance, allowed our customers to get more out of automation with very little effort. 

  • With self-learning, cognitive bots that could apply machine learning and artificial intelligence along with natural language processing, automating complicated processes (that were previously considered too ineffective to be automated) became a real possibility.

  • Capturing unstructured data, cultivating a deeper understanding of processes, and taking human-like decisions was no longer robotic science fiction but hard-core fact. 

  • Customers now had the leeway to reduce the manual intervention that came out of rule-based automation — where each change or exception in a process required human effort to fix issues or train the bot.

  • The need to draw elaborate, custom frameworks based on industry, domain or vertical was also eliminated through domain-specific packages. JiffyRPA can be readily leveraged across several domains — Finance and Accounting, Human Resources, IT Services, Business Process Management and more.

  • More so, enterprises could easily implement a continuous automation framework with the advancedprocess automation, reporting and analytics capabilities that JiffyRPA provided. The Jiffy analytics engine calls out bottlenecks in the process and highlights opportunities that could enhance the outcome of automation. This meant customers could react faster to opportunities and scale operations as required – all while keeping maintenance costs to a minimum.

In such a demanding environment, automation needs to be able to set the pace at which it seamlessly fits into a business—right from reducing turn-around-times and increasing productivity, to becoming a strategic value-add by complementing your business objectives. This is where JiffyRPA’s cognitive automation works — and amazingly well at that.

Automation + Intelligence – The only way forward

Sudhir Sen, Products Head of JiffyRPA asks, while automation is the poised to be the driver of change, has it truly embraced its potential as we see it today?

 

option3-jiffy@automation

Image by sdecoret/shutterstock

Enterprise transformation has been accelerated with the rising capabilities of automation. While it is the poised to be the driver of change, has it truly embraced its potential as we see it today? Its benefits are evident and results have been promising, and intelligent automation is vaunted as the next big step . Yet the pace in which Intelligent/Cognitive automation is being adopted is rather slow. My conversations with senior leaders across different industries and functions have changed over the last 12 months. Intelligent Automation was unheard of/new to the picture back then, and now it’s quite the flavor, but still remains a grey area. Decision makers explore long term benefits with their automation journey planned in different stages of execution, with cognitive capabilities becoming a requirement only after they’ve established a degree of familiarity with current automation.

Let me throw more light into why cognitive automation makes sense, and why it needs a Day 1 approach.

Cognitive is beyond rules

With enhancement, businesses can automate the decision-making processes by itself. This is where most businesses struggle to remove the manual dependency on repetitive processes. These tasks are considered complicated and rarely given a thought to be automated. If the software bot can train and learn on itself, and semantically understand the process and data and understand from history what the next course of action is to be taken, it will free up more time for staff and increase the levels of automation for business.

Exceptions always require interventions, and with self-learning capabilities, bots can minimise these interventions, thereby increasing process efficiency and enhancing accuracy with their cognitive capabilities to assist decision making. The right automation solution will even help you bring down costs from investing in unnecessary licenses for bot farms. It will even decide based on load and the availability of bots how to distribute work and get things running without pausing for allocation of resources.

Let’s take the example of a common challenge of invoice processing and compare two scenarios.

Typically, with rule-based automation:

  1. Bot designers manually define templates for each supplier graphically
  2. Bot reads new incoming invoice and identifies the template to be applied
  3. Bot extracts information based on template and loads into ERP
  4. Bot tags invoice for manual processing if a suitable template could not be identified
  5. Operator manually loads the invoice to ERP
  6. Bot designer manually adds the templates for invoices which could not be read

See the point? Each time a new invoice enters the system and the bot is not able to read it, the designer needs to train the bot with the new template.

The same process when executed by a cognitive bot:

  1. Bot designer extracts historic invoices and ERP data
  2. Bot reads the invoices and generate Machine Learning model
  3. Bot reads new incoming invoice and applies ML model to extract invoice details
  4. Bot loads the data into ERP
  5. Bot tags invoice for manual processing if it cannot read with sufficient confidence level
  6. Operator manually loads the invoice to ERP
  7. Bot self learns how operator read the invoice and handles it the next time

The difference is visible. It’s equivalent to how we think, without rules. Cognitive automation understands the context and content and handles changes intuitively.

Get your worth out of automation

Cognitive automation reduces manual effort to unprecedented levels once RPA is implemented. It can handle a wide range of complexities and unstructured data within a short span of time, and give you higher automation levels without restricting to rules/creating new templates each time the process encounters and exception. It is not something to be kept for later stages as an afterthought to existing automation, and needs to be a key strategic asset while looking for automating from day 1.

It is important for enterprises to educate themselves on revisiting their approach to automation, when the possibilities are endless with this platform that has real cognitive capabilities.

Source :  Artificial Intelligence 

Challenges of implementing RPA in Logistics

Each process workflow is unique to how it’s practiced in a logistics company. This requires a customized approach to integrate all these different processes and automate them end-to-end.The biggest challenge is the lack of pre-built automation solutions that can seamlessly fit across any workflows.

Sudhir Sen -Co Founder -Option3

Logistics has always been a slow adopter of new technology, primarily because of the complexity of operations involved and the reluctance to shake up things when it’s seemingly in order and manageable. However, with the opportunities through technology to accelerate processes increasing by the day, it has injected a much-needed vitality to logistics in the recent years. It’s still lagging when compared to other industries, but the pace is picking up with automation managing to effectively complement both the functional and transactional processes.

Sudhir Sen, Co-Founder, Option3

The biggest problem plaguing adoption is the lack of pre-built automation solutions that can seamlessly fit any workflows and start automating processes. This is because processes may be similar in function yet so different in execution and operation. Each process workflow is unique to how it’s practised in a logistics company, and this requires a customized approach to integrate all these different processes and automate them end-to-end.

The complexity of automation is evident across different logistic functions – from Order Management and Processing, Procurement, Distribution, Warehouse Management etc, as they use different software to manage large volumes of processes. There is a wide range of software products that assist in operations with custom solutions being developed for each logistics operator. This directly leads to higher costs of automation.

There are other challenges associated with logistics automation, like the huge data volumes and quick responses expected in the process. New age RPA (robotic process automation) solutions can even monitor real-time inventory, reorder products based on optimal inventory and bring in advanced cognitive capabilities. The robots are even designed to understand business data patterns and make decisions on it.

Marrying software with hardware automation can address another major gap plaguing the logistics industry. Established players have already implemented robots in managing warehouses and distribution, yet there’s still a lot of progress to be made in mapping it to software that completes the cycle in terms of order management, customer service, billing, operations and more. The capability to take inputs from hardware robots and execute actions using software bots are evolving.

In a highly demanding operation such as logistics where turn-around times are critical to businesses without any trade-off in quality and efficiency, automation can help operators to achieve the last mile.

Source: challenges implementing rpa logistics

Scaling from tactical RPA to strategic RPA with Cognitive Automation

Enterprises in the early stages of their automation journey are still apprehensive about implementing Robotic Process Automation (RPA); those who are implementing RPA are struggling to remove human intervention. Many a time, RPA is bottled down as an immediate solution, which is far from true.

Sudhir Sen -Co Founder -Option3

Robotic Process Automation (RPA) is gaining momentum by the day, all thanks to the advancements in automation technology over the last few years. Businesses of all sizes— from enterprises to start-ups— have realized the value that automation brings. We now see automation solutions delivering high impact, focused results across various domains— from shipping and logistics to software development and ecommerce. And all businesses realize the need for RPA across their different functions— Finance, Payments, HR, IT, Operations, and more.

Sudhir Sen, Co-Founder, Option3

The RPA market is predicted to grow at 70-90% and become a $600-800M market by 2018 (Everest Group Report) and is expected to grow to $2.9 Billion by 2021 (Forrester).

What we have observed over the past couple of years is that enterprises in the early stages of their automation journey are still apprehensive about implementing the RPA and the changes that come with it, while enterprises who have advanced implementing RPA across their journey are still struggling to remove human intervention in some of the complex processes—thereby not really getting the best out of their RPA strategy.

Added to this is all the talk of Artificial Intelligence, Machine Learning, NLP and what not, to confuse enterprises further when they are still struggling to get to grips over RPA. It’s still murky waters from a decision-making standpoint, as there is no clear structure to how it may be approached, and this has bottled RPA down to being treated as a mere tactical, immediate solution. From where we stand, this is far from true. Let us elaborate.

Cognitive or AI is the current flavour with anyone engaged with RPA right now. But all the talk still masks the fact that it’s just a lot of rule-based automation with an AI layer to the bots to pass it off as cognitive. True cognitive capabilities are supposed to replicate how a human would be engaged during the process—this involves self-learning, decision-making, and the ability to process unstructured data into rational information.

This gives enterprises the capability to plan for the future without worrying about changes or transitioning of processes or getting bogged down with operational bottlenecks, and can truly enable them to engage a highly advanced digital workforce to complement their business.

New entrants to RPA still treat it as an immediate fix because it requires minimal effort to implement. Define a process, and automate the workflow. Save time and add more efficiency to the process, to achieve better and faster results. However, when you want to scale up and add more processes to automate, you may hit a roadblock due to localization, complexities, unexpected data formats and architectural challenges. This is precisely why it’s still considered as a tactical productivity enabler.

However, with the right RPA solution—one that is self-learning, scalable to handle any type and volume of processes and can automate the most complicated tasks easily, with the ability to process and provide insights into unstructured data that it deals with—enterprises can now engage more on optimizing processes rather that getting stuck in RPA operations. This takes the benefits from RPA to a different level altogether.

Consultants, service integrators, and solution providers for the most part only advise businesses about considering cognitive automation once they have attained a familiarity with RPA, and this advice is consumed without an afterthought. Nothing could be farther from the truth. That is one of the reasons RPA 2.0 is declared to be the next gen RPA. But, enterprises need to define what they need to achieve out of cognitive automation. The very fact that they are still considering this means that rule-based automation or traditional RPA needs to evolve.

While efficiency and speed can be highly improved with traditional RPA, cognitive RPA—with the ability to engage with processes with self-learning capabilities as much as a human would —is where RPA is going to define itself in the coming years. With the right information at hand to effectively optimize processes, enterprises stand to gain a lot by considering cognitive RPA as a differentiator and a strategic asset to drive their business.

It’s never late to future-proof your RPA. Here are the key considerations to make your RPA to an iRPA (Intelligent RPA):

• Continuous learning—Machine Learning models should be trained frequently to match the decision-making frequency depending on the diversity of the input data.

• Robust Decision making—Enabling your RPA to take decisions on input data that was never encountered before.

• Taking your OCR to next level—Making your OCR intelligent is key to making your RPA self-sustained.

Source: Scaling tactical rpa strategic rpa cognitive automation

Cognitive Automation Delivering More with Intelligence and Sophistication: Option3

Business processing now requires more computational and processing capacity with the volume of data that can continues to grow

Payeli Ghosh (B) and Shreyas PC (F) - co-founders, Option3.
Payeli Ghosh (B) and Shreyas PC (F), co-founders, Option3

Cognitive automation is delivering increased efficiency, and speed and reduced costs, and more with intelligence and sophistication. It can learn on its own how to manage changes within the system, make sense of unstructured data, take decisions like a human would, analyze the processes, and data it interacts, with offers suggestions to optimize the processes further.

Payeli Ghosh and Shreyas PC, co-founders, Option3, speak more about the cognitive robotic process automation. Excerpts:

BW CIO: What are the major driving forces that tend to increase the demand for cognitive robotic process automation during the forecast period 2018-2026?

Shreyas PC: There are many factors.

The pursuit for excellence
While traditional rule-based automation has been delivering results with increased efficiency, and speed and reduced costs, cognitive automation goes beyond that by delivering all of these and more with intelligence and sophistication.

It means, it can now learn on its own how to manage changes within the system, make sense of unstructured data, take decisions like a human would, analyze the processes and data it interacts with offers suggestions to optimize the processes further. Cognitive capabilities add more dynamism to automation, and businesses stand a lot to gain from it.

For example, if there’s a change in a process in the form of data input, or an additional step for validation, rule based RPA solutions would need to be programmed in order to facilitate automation. With cognitive RPA, the bot analyses the change and tries to understand the context and content all by its own and then acts accordingly due its self-learning and decision-making capabilities.

Automating complicated processes
Business leaders start looking at cognitive automation when rule-based automation fails in simplifying processes that are complex and dynamic in nature, such as non-standard exceptions, and unstructured data. This requires logical decision-making capabilities in the RPA platform which is possible only through cognitive RPA.

For example, when it comes to a decision to match an invoice to a goods receipt, in most cases there are no pre-determined rules, and requires a correlation capability with an understanding of the system in place. Cognitive RPA helps to match it by checking the records and understanding if the goods have been received by the merchant.

Big Data
As businesses grow, so does the data. Business processing now requires more computational and processing capacity with the volume of data that can continues to grow. Cognitive RPA, while minimizing the burden of the repetitive tasks, can also store the data which can be used for its learning. This makes it powerful enough to hold all this knowledge to understand patterns and trends and create reports and metrics for businesses to understand and achieve their goals.

BW CIO: Could automation force a re-organization of the industry?

Payeli Ghosh: It’s similar to how tractors came in and revolutionized agriculture. The outcomes have not changed, but farmers adopted technology to increase efficiency and quality of farming. It enabled them to focus more on other areas such as sourcing, distribution, and sales.

Automation is similar in many ways, as people will start focusing more strategic activities, and result in them moving up the value chain. Automation will create new roles and open new opportunities for industries, and we can also see teams working in tandem with virtual agents to achieve outcomes in a smoother way.

BW CIO: What are the major challenges inhibiting the growth of the global cognitive robotic process automation market?

Payeli Ghosh: Most organizations are pursuing RPA in varying degrees, but when it comes to cognitive RPA, they don’t know where to start. They are usually advised to achieve a degree of familiarity with rule-based or standardized RPA before venturing into cognitive or Intelligent Automation. It is not the right approach as they would miss out on a lot of efficiency that would be gained through a cognitive solution in the initial stages.

Another case that has been observed is organizations hitting the reset button and opting for a fresh start with cognitive RPA when they fail to realize the results through rule based automation. Had a holistic approach been considered, a lot of these challenges could have been overcome.

Another challenge is the common perception among decision-makers and influencers that cognitive RPA solutions are difficult to build and integrate to their business. It’s true in some ways, as a cognitive solution would take more time to build than a rule based automation platform, but with the right RPA product, it is much more faster and a lot more efficient.

For example, to create a standard cognitive robot does not take more than 6 weeks with JiffyRPA and if it is based on any of the predefined solutions, it does not take more than 2 weeks to bring up a cognitive bot.

BW CIO: What are the types (services and platforms) of cognitive robotic process automation market that will dominate in the coming years?

Shreyas PC: RPA solutions integrating Natural Language Processing and AI capabilities will become standardized in the coming years. Robots will be more human-like in nature where they would be able to interpret and respond to situations on their own without any manual intervention required.

It will continue to be a partnership between solution providers and service integrators to deliver results for enterprises.

BW CIO: What is the total market share of the multiple industries (finance and banking, insurance, healthcare, telecom and IT services, and others) in the global cognitive robotic process automation market?

Payeli Ghosh: Riding on the successes of RPA and cognitive capabilities that are rapidly increasing, there is a rise in demand across industries and functions looking into cognitive RPA. According a 2017 report by Shared Services and Outsourcing Network, Intelligent Process Automation market is expected to hit US$ 1.2 bn by 2021.

BW CIO: What are the major opportunities that cognitive robotic process automation companies foresee?

Payeli Ghosh: As elaborated earlier, cognitive automation comes in place to automate complicated tasks that require a level of intelligence to replicate human-like decision making. Opportunities are many, as businesses look to invest more in simplifying complex tasks to ease the burden on their employees and increase their productivity while accelerating business outcomes.

This includes reading and processing documents, forms and invoices which are in different formats, processing and preparing contracts, autocorrecting unstructured data and so on.

Source: Cognitive Automation

Request Demo

Bitnami