Cognitive Automation — Going beyond Rule-based RPA

People And Robots Modern Human And Artificial Intelligence Futuristic Mechanism Technology Vector Illustration
In recent times, organizations across the world from various industry sectors are pushing themselves to become “Digitally Native” by adopting “Digital Transformation” as a foundational pillar for the future starting with Robotic Process Automation (RPA). And, the primary objectives for the most of them are to achieve speed, accuracy, and reduction in headcounts. The convergence of RPAartificial intelligence (AI)machine learning (ML)natural language processing (NLP), and cognitive platforms is potentially so disruptive that Klaus Schwab, founder of the World Economic Forum, calls it the“Fourth Industrial Revolution.” At the same time, there is a good share of apocalyptic warnings from various quarters that the advancement of automation (cognitive form) in the workplace will create a “dystopian society”. To nullify such warnings, Karen Lachtanski aptly wrote, -“If an argument is to be made against digital transformation, it is that the divide between high-level skills and low-level skills will become wider, with little or no middle ground.” In short, she clearly states that there is no time to dwell on it. What going to set the future workforce apart is what we as a part of that workforce are willing to do about it i.e, learn and evolve (or simply perish).

What is Robotic Process Automation (RPA)?

RPA is a basically a software tool that can automate routine tasks/sub-tasks in structured mannered by eliminating human activities such as “copying and pasting data between multiple applications” so that functional/cross-functional teams can focus on more value-adding activities.

In December 2017 survey by Deloitte“53% of the respondents have already embarked on the RPA journey and a further 19% of respondents plan to adopt RPA in the next two years”. If adoption continues at its current level, RPA will achieve near-universal adoption within the next five years. One reason for such prediction is that RPA has become advanced enough to take over the mundane tasks; prior to that, the technology wasn’t quite there.

Let us accept, broader the automation spectrum, more the elimination of manual processes. For organizations to become digitally native it’s very much important that an RPA technology should be designed and deployed as an ideal tool to connect multiple legacy systems rapidly and seamlessly such as Big Data, Internet of Things (IoT), cloud, etc. Eventually, it should become a critical part of their value proposition just not for the internal operations but also for the front and back-office functions.

The Addition of “Cognitive Intelligence” to create Cognitive Automation

While RPA is expected to act as a first step in the adoption of automation, the rise of new cognitive technologies (which can mimic human intelligence and judgment) is expected to increasingly drive automation by matching the current wave of “Digital Transformation” with the application of AI. In fact, cognitive technologies can be considered as a subset of AI, further grouped into capabilities such as ML, NLP with semantic analysis, machine vision, speech recognition, emotion recognition with sentiment analysis, and optical character recognition.

On August 15, 2018, Deloitte and NICE launched a white paper – “The Future of Operations — Moving Beyond Process Automation” which meticulously covered a  futuristic self-service banking scenario that utilizes a myriad of new generation cognitive tools to stay ahead. The paper duly explained the concept called Robotic and Cognitive Automation (R&CA) with a holistic and rich perspective on “how to practically assess and tackle the next technological revolution in artificial intelligence and cognitive automation”. Unlike the RPA, given their probabilistic nature, cognitive technologies need to continuously learn from their past actions and evolve more accurate algorithms.

One of the biggest constraints of RPA is that it needs structured data in the form of a spreadsheet, a web form or a database for the robots to work flawlessly. Hence the need for cognitive intelligence (driven by ML/NLP) arises to deal with the unstructured, or semi-structured data and transform it into a structured form, which can then be later processed by the robots.

WorkFusion’s Smart Process Automation (SPA) is one of the classic examples, which is, in turn, paired with RPA to learn from the humans it supports. Using ML-driven data capturing tools, inbuilt quality control, and algorithmic training capabilities, bots shadow human actions and judgment calls to learn routine decision-making processes.

Kindly do note, many advertised AI-powered RPA solutions often turn out to be a basic extension towards ML (Not purely driven by ML as such). Basically, such extensions are quite useful, but they are based on recognition patterns, i.e., having a rule-based dependency. Before selecting any such solution, a due diligence is recommended.

No matter how lucrative cognitive automation seems to be, the first-mover‘s pursuit in this space may invite risk. The best possible strategy is to test run a set of pilot programs and then evaluate for a smoother downstream implementation. A set of proven pilot results can easily help an organization to formulate a long-term strategy.

The Road Ahead

In a newly published working paper by Lukas Schlogl and Andy Sumner from the think tank, the Center for Global Development (CGD) explained the potential effects of robotics and AI on global labor markets. When automation is used to augment human management, traditional organizational orthodoxies, such as about spans of control, can be challenged. The paper says it’s impossible to know exactly how many jobs will be destroyed or disrupted by new technology. But, authors add, “it’s fairly certain there are going to be significant effects — especially in developing economies, where the labor market is skewed toward work that requires the sort of routine, manual labor that’s so susceptible to automation”.

As in the past, technology will not be a purely destructive force like the introduction of Automated Teller Machines (ATMs) pushed down the branch-wise headcounts but at the same time banks got an opportunity to open more branches at the distant corners. In this particular case, new jobs will be created; existing roles will be redefined, and workers will have the opportunity to switch careers. But, the challenge to this generation will be in managing the transition as the individuals who need to retrain for new careers won’t be the young, but middle-aged professionals.

And from the government’s end, policy-makers should embrace the opportunity for their economies to benefit from the implementation of cognitive automation and boosting productivity across. To achieve that, they should put in place well-defined policies (flexible, not rigid) to encourage investment and offer market incentives to encourage continued progress and innovation. At the same time, they must evolve and innovate policies (keeping pace with time and evolving technologies) that help current and future workforces adapt to the impact on their respective employment demographics. The dawn of “automation age” has already arrived and it needs an extensive level of social re-engineering which must include revamping the education and training systems, creating substantial income support and pre-defined safety nets, as well as a necessary transitional support for those dislocated or about to be dislocated.

About the Author:

Rahul Guhathakurta (ORCID: 0000-0002-6400-6423)

Cite this Article:

Guhathakurta, R., “Cognitive Automation — Going beyond Rule-based RPA” IndraStra Global, Vol. 04, Issue No: 9 (2018), 0006, http://www.indrastra.com/2018/09/Cognitive-Automation-004-09-2018-0006.html | ISSN 2381-3652

Read More

Decoding Uber’s Rating System for Drivers and Riders

Image Attribute: Uber ride Bogota (10277864666) / Source: Wikimedia Commons

By Rahul Guhathakurta

Being one of the most frequent users of Uber app, it has always boggled my mind how we as a passenger get rated by the drivers and what kind of impact it does have on the services and prices pertaining to us. Finally, the secret has been spilled out with the recent update of Uber app which has a noticeable change for Uber riders – an addition of a rating below the user’s name in the menu of the app. When a rider tap on the menu button, he or she can see his/her profile picture, his/her name and a score that represents the average rating the rider has received from drivers they have ridden with.

How does it work for drivers?

According to Uber’s official statement, an average rating of 4.8 or higher for a driver is considered to be “outstanding” and the company itself encourages the drivers to aim for the same. But, the ratings do have other demarcations which are as follows; If a driver’s average rating is above a 4.6, the profile is considered to have a “good standing”, but there may be a few ways a driver could provide an even better experience for their riders. Drivers are encouraged to maintain at least a 4.6 average over their most recent 100 trips. If the average rating is between a 4.3 and a 4.6 after first 25 trips, drivers will need to improve their average rating to a 4.6 over the next 25 trips. If the rating over the most recent 100 trips is below a 4.6, the driver’s profile may be at risk of deactivation.

If a driver’s average rating is above a 4.6, the profile is considered to have a “good standing”, but there may be a few ways a driver could provide an even better experience for their riders. Drivers are encouraged to maintain at least a 4.6 average over their most recent 100 trips. If the average rating is between a 4.3 and a 4.6 after first 25 trips, drivers will need to improve their average rating to a 4.6 over the next 25 trips. If the rating over the most recent 100 trips is below a 4.6, the driver’s profile may be at risk of deactivation.

In the case of UberSELECT, driver’s average lifetime rating must stay above 4.7. If the rating falls below 4.7, the driver will not be able to receive UberSELECT trips until his/her ratings improve. Also, ratings on high surge trips will also not be taken into consideration when determining whether driver’s profile is in good standing.

How does it work for riders?

According to an anonymous Uber employee on Quora – In the United States, there is a huge difference when “chillaxed” West Coast cities like Los Angeles and San Francisco see average ratings for riders and drivers that are higher than aggregated ratings given by East Coast riders, who are more likely to reserve their 5-star ratings for truly exceptional experiences. This applies to both rider ratings and driver ratings.

Without divulging company information, the same anonymous Uber employee said – “The best answer I can give is to compare your rating against the ratings of your local drivers. If good drivers in your area have an Uber rating of 4.7 out of 5, this likely translates to a good passenger rating as well.”

The employee signed off by giving an advice which needs to be kept in mind that small sample sizes will throw off rider’s rating, often skewing it lower than it would be otherwise. For example, some drivers default to rating each passenger 3 stars, the same way that some riders rate their drivers 3 stars by default. If a rider had one or two drivers that do this, it will drop the average rating until the rider had taken enough trips to smooth out his/her score. Unless a rider had taken 20+ trips, the passenger rating has probably not settled to its actual value.

“Parallel Handshake”

It’s the first time Uber has brought in transparency between Riders and the Drivers through “Parallel Handshake”. In fact, the rider rating is sort of the driver’s communication tool to get its prospective passengers ratings based on later’s behavioral aspect. Low ratings usually indicate someone who is rude, a backseat driver, doesn’t tip, requests a lot of things, makes multiple stops or just isn’t cooperative with the driver. So a low rating will either make rider’s pickup request longer or rider may not be able to get a ride at the time of utmost need. Besides that, Uber could deactivate rider’s account if the rating gets too low.

Conclusion:

Let’s be very clear, the Uber ratings are generally reported as averages, and neither riders nor drivers will see the individual rating left for a particular trip. It is recommended that as a rider one should maintain and display the utmost level of decency and friendliness to reap the benefits of Uber ecosystem. Do note, tipping isn’t required. However, if a rider is looking to boost his/her rating, tipping should certainly help at the certain level. Beyond it, it’s all about how rider behaves. Period!

About the Author:

Rahul Guhathakurta, Brand Principal at AutoKrew and Founder/Publisher, IndraStra Global. He is a seasoned retail and supply chain consultant with cross-industry experience spanning 3 continents. He regularly tweets at @rahulogy.

This article was originally published at CarKrew.com

Read More

Blockchain in Automotive Domain

Blockchain in Automotive

Transparency, Interoperability, & Scalability

 By Rahul Guhathakurta

The “Blockchain” is a revolutionary database that does away with the weaknesses of traditional solutions for storing big data. It provides a transparent record of the entire business network, allowing buyers and sellers of each vehicle to track where the vehicle is in its lifecycle. The blockchain as a technology is more of a ledger recording “agreements.” It is a system which contains a continuously growing list of records, called blocks, which are linked and secured using blockchain-based Public Key Infrastructure (PKI) encryption. It also provides a complete history of “deals” made between two or more parties, in which the record cannot be altered retroactively without the alteration of all subsequent blocks and the collusion of the network.

Upstream Application

The advantages of the blockchain are valuable, to say the least. Automotive manufacturers can partner with a blockchain service provider to create a unique ledger among its network of original equipment manufacturers (OEMs) as well. It can help them to address the issues related to parts quality and the cases of fictitious pickups to strengthen its logistics monitoring and control. Like, one can cut tracking time of a particular shipment or series of shipments from several days to seconds. With improved traceability, both automotive manufacturers and OEMs can ensure the authenticity of parts delivery at assembly level milk runs.

Software-based Manufacturing

Software-based manufacturing based on blockchain can easily increase the manufacturing productivity and quality, significantly reduce the expenses for tracking in regards to warranty, maintenance and recycling purposes. Factors related to extra inspections about the product fabrication, version management, and recalls/callbacks would also, be simplified.

For an example, a unique sensor can be added to each and every parts pallet before it is shipped from original equipment supplier (OES)-end and tracks its real-time status and location, to ensure that the shipment complies with the expected time of arrival (ETA) at the assembly dock. This entire works in conjunction with the Internet of Things (IoT) to form an effective anti-counterfeit strategy by using countermeasure technologies based on blockchain principles, where each supply chain partner proactively takes part in updating the status (attributes) of the item as it traverses from point of sourcing to point of manufacture to point of sale. The whole concept demands an extensive and exhaustive level of cooperation among automotive manufacturers, OEMs, software developers and cybersecurity firms at a scale that has never been achieved before.

Downstream Application

Vehicular Maintenance

On July 25, 2017, Renault, the French automaker announced a pilot project to digitized its car maintenance program, which uses blockchain as a shared ledger to log all car repair and maintenance history in one place. According to Renault, currently, the information about a vehicle’s maintenance history is kept by a range of sources, including repair shops and dealerships, making it harder to keep track of new changes. By contrast, the blockchain-based digital maintenance log prototype puts all of this information in one place. The next pilot, it says, will focus on vehicle-based microtransactions – essential to integrating the IoT with the exchange of value.

Automotive Finance & Vehicle Leasing

Automotive financing varies somewhat by the nature of the transaction, typically it includes a host of verification steps to which blockchain could be applied for efficiency gain over the traditionally cumbersome processes like customer bank validation along multiple phases of transaction set up in compliance with know your customer (KYC), loan approval; review of multiple documents sourced from different locations; scoring and classifying risk; archiving of reviewed documents; etc. A blockchain-enabled smart contracts, which will allow these files to automatically negotiate payment on a new car lease and other terms and conditions with the prospective leaseholder, without the need for a middleman. It will also ensure the execution of secure crypto-payments to the necessary parties.

A dedicated blockchain system can also provide most-needed transparency of information about a vehicle’s real wear and tear would help the auto finance provider to more accurately gauge the residual value of the vehicle as the lease approached its end of contract date. For example, an On-board Diagnostic-II (OBD-II) device connected vehicle over a defined mobile network, would enable capturing of data like driving behaviour events (mileage, hard break threshold exceeded), safety events (airbag deployed, part replacement warning), service events (annual service, part replaced), etc., and get sent to a shared ledger that all parties had access to, including the owner/leaseholder. This, in turn, would enable the auto finance provider to achieve a higher price at subsequent onward sale than would otherwise be possible.

Fleet Tracking

The same fundamental of OBD-II devices can be deployed, riding over a blockchain infrastructure, which, in turn, can enable fleet companies to push & pull OBD-II messages, fetch the gyroscope inclination, along with GPS position of a vehicle or a fleet in real-time. Further, it can be enhanced and integrated with Electronic Logging Devices (ELD) in compliance with the Federal Motor Carrier Safety Administration (FMCSA) mandate, providing an all-in-one tool that streamlines every facet of a truck driver’s job. A cloud-based, blockchain-driven fleet tracking can solve the key issues related to drivers, dispatchers, and fleet owners often face with the best hardware and software available in the market, while also fetching a reasonable rate of return on investment (ROI) in near future.

Conclusion: An Integrated Application

Overall, a dedicated multi-tier interconnected blockchain platform based on the fundamentals of scalability and interoperability can benefit many stakeholders, like – a shared ledger – between automotive manufacturers, automotive dealerships, regulators, auto finance-cum-insurance companies, vehicle leasing companies, buyers, sellers and even garages, providing a higher degree transparency and trust in all kind of vehicular transactions, preventing disputes and lowering the overall cost of maintenance and services by tracking ownership, sale, and accident history. And, at the same time, it could significantly streamline processes, especially those that rely on regulatory and compliance approvals. The blockchain is all about bringing in transparency and efficiency into the existing systems which are running the upstream and downstream supply chains and making them more proactive and predictive.

About the Author:

Rahul Guhathakurta, Brand Principal at AutoKrew and Founder/Publisher, IndraStra Global. He is a seasoned retail and supply chain consultant with cross-industry experience spanning 3 continents. He regularly tweets at @rahulogy.

Cite this Article:

Guhathakurta, R. “Blockchain in Automotive Domain: Transparency, Interoperability, & Scalability”, IndraStra Global Vol. 004, Issue No: 03 (2018) 0024, http://www.indrastra.com/2018/03/Blockchain-in-Automotive-Domain-004-03-2018-0024.html | ISSN 2381-3652 | DOI: 10.5281/zenodo.1197047

Read More
Subscribe to our Newsletter
  • Email ID:
  • Name:
  • Company: