Despite highly mature IT landscapes of German insurance companies, there is a massive need to catch up when it comes to robotics. Large volumes in dark processing and many processes that still need to be handled manually make the industry tailor-made for the use of Robotic Process Automation (RPA). The use of digital robots in the fully automated processing of structured business processes promises insurers time savings of up to 90% and cost reductions of up to 80.

by Thomas Sellner, Head of Software Development, Almato

A vacation is booked quickly. On the Internet with just a few clicks. If only the corresponding insurance confirmation could be obtained just as quickly. But the procedure is like fighting with a bureaucratic time-consuming monster.

Almost bizarre: The tiresome battle is not only waged on the customer side, but also on the company side."

Again, employees struggle through various software systems – from web applications to CRM and industry-specific solutions – on their way to confirming insurance, tediously picking out the different forms required depending on the destination country and reason for travel, and having to put everything together by hand before the documents finally make it into the mail. That's a lot of work for employees, which means annoying – and usually incomprehensibly long – waiting times for customers.

Efficiency advantages through digital robots

And it could all be done in just a few clicks. Robot-based process automation makes it possible. The entire process of sending an insurance confirmation, such as for a foreign insurance policy, can also be handled via Robotic Process Automation, or RPA for short. RPA is the fully automated processing of structured work processes by software, the digital robots. These robots work like a human on the front end, only much faster and more accurately, while documenting everything at the same time.

Companies could reduce processing time for tasks by 80 to 90%."

Because robots, unlike humans, don't make mistakes, they not only make for happier customers, but more importantly help insurance companies meet stringent regulatory requirements for data security and quality and auditability.

While in Anglo-American countries, companies from the financial and insurance sectors have long been using RPA, the use of robotics in this country is still in its infancy. Decision makers find the new technology too abstract. They lack reliable figures and experience values with RPA. However, in view of the omnipresent cost pressure, robots are increasingly coming into focus here as well, and not only in the context of outsourcing negotiations.

Workflow support

RPA solutions can be easily integrated into existing IT landscapes; complex system integrations are not necessary. RPA can be used wherever structured processes are fundamental: Whether for insurance confirmations, online new registrations, processing of contracts in the event of the death of the insured, changes to address data, direct debits and standing orders – everything based on business rules can be handled via robot-based process automation. The robots record the data, calculate the (new) insurance amount if necessary, change the data in all systems and automatically send the (updated) contract to the customer or make it available as a download. Even when the insurance claim occurs, the claim process can be carried out via RPA: Standardized appraisal and repair orders are recorded and then automatically processed further by the robots, confirmation email to the insured included.

Thomas Sellner is Head of Software Development at Almato. The interaction of people and technology has been one of the focal points of interest for many years now.

After studying media and communication informatics at Reutlingen University, he worked at the Fraunhofer Institute IA as a research assistant for the Human-Computer Interaction Competence Center. Thomas Sellner has been working as a software consultant at Almato since 2012 and, as head of software development, has been responsible for the technical management of all software developers since August 2014.

Non-standard processes may have to be outsourced to an employee, but RPA helps here too by having the robots bundle all the necessary documents so that the employee doesn't have to spend time on lengthy searches, but "only" has to make the decision.

RPA is not only applicable in product as well as customer data management by always automatically updating data across all systems, but also in mass processing of billing, claims payments and recourses. Automation via RPA supports compliance rules, as well as reporting, and also helps with the introduction of new products and corresponding campaigns.

RPA as rule-based software is flexible

For insurance companies, against the backdrop of the large volumes in dark processing typical of the industry and the many manual processes, this is tantamount to a considerable optimization potential that only needs to be leveraged. Because, let me assure you, the use of robotics provides significant cost benefits, efficiency gains and more satisfied customers due to faster turnaround time. 

The post Dark processing: robotic process automation for insurers – offloading clearly defined tasks appeared first on education.telefony-taksi.ru.

Robotics applications and artificial intelligence can help banks and insurance companies efficiently meet the growing needs of customers. However, while rule-based bots are already being discussed or even tested in almost all major companies in the financial sector, there is still a long way to go before "real" artificial intelligence can add value for the institution.

By Dr. Hendrik Schreiber, SQS Software Quality Systems

There is hardly any other topic in which the banking industry is currently expecting as much efficiency gain as from the automation of processes. Whether it's robo-advisors and bots in consulting or automation tools in the back end that evaluate large volumes of data or reconcile databases – there is potential for cost savings and reducing the workload of employees at many points in the workflows. Especially for routine tasks.

However, establishing automated processes in the banking world is complex and not everything that is sold under the buzzword "artificial intelligence" deserves this label."

Indeed, the term artificial intelligence is currently being used somewhat inflationarily for many things that, on closer inspection, could at best pass for advanced automation. Robotic Process Automation (RPA) can be used to automate very similar processes in the back office on the basis of rules, and these processes can also take place across systems. For example, a bank can thus convert certain data sets in the same way based on rules or process them according to certain specifications.

For SQS, Robotic Process Automation (RPA) is an important business area with increasing significance. The company offers the topic holistically from a single source – from business analysis and process identification to implementation and quality assurance as well as rollout. In principle, any Windows application can be automated – optionally also Web services. SQS uses the Blue Prism tool, which is intended to support the automation of manual banking and finance processes in the future. The technology behind it does not require any special IT infrastructure: The software robot runs on a standard Windows platform and can therefore be used in all Windows-based environments. In principle, any Windows application that can be accessed from a Windows workstation can be automated.

Artificial intelligence as the supreme discipline of automation

Automation based on artificial intelligence, on the other hand, requires a learning system that not only has a huge pool of data at its disposal, but also learns successively on the basis of runs and completed cases. The data structures here are considerably more complex than in the other cases. Decisions can still be made in real time due to increasingly powerful IT resources.

This increase in complexity is also reflected in the higher effort required for testing. In addition to the usual test procedures, new ways must be found to deal with self-learning systems. One possible way is crowd testing, for example when testing chatbot-based applications. When testing self-learning systems, the data used to teach artificial intelligence must also be carefully selected from a statistical perspective. This is necessary to ensure that the test does adequate justice to the later reality.

Savings potential as the most important incentive

First, a pilot project is usually carried out to show the customer the potential and demonstrate the added value – and to show him that and how such an application works. Together with the customer, the company carries out a process analysis to select the processes for which savings potential can be expected on the basis of volume and automation capability. Because, despite all the other advantages such as 24/7 operation and faster end-to-end processing, this is actually always in the foreground. One or two suitable processes are then selected from a list of possible processes.

Dr. Hendrik Schreiber already worked as an application developer for optimizing customer service while studying computer science at the University of Paderborn. Following his studies, he did his PhD on the topic of test effort estimation at the s-lab – Software Quality Lab at the University of Paderborn from 2010 to 2015. Since 2015, he has been working as a test manager for SQS Software Quality Systems . Since 2016, he has been leading a team with a focus on digital banking in the role of Practice Team Lead.

Indeed, there should not be too many at the beginning, because such a pilot project, in addition to the actual automation, should also serve to get to know the company's IT infrastructure and the quality of the data, and to adjust to the customer's workflows. Because companies themselves usually know their own processes quite well, but an unbiased external view often brings a gain in knowledge. One hurdle, for example, can be different databases and sources of information coming together. The important thing is that the data must be available in a machine-readable form. Text in images, for example, must first be recognized and correctly converted using OCR processes. This may require further transformation steps into suitable data structures.

Change process: Bringing MA along!

But accompanying the change process among employees is also an important element in the introduction of automation. Finally, robots, whether in hardware or software, have the potential to replace many employees and you have to consider ..

… how to honestly and credibly signal to the employee that the robot is a tool that facilitates and changes repetitive, boring work, but in no way makes the employee superfluous."

Verbal communication as a special challenge

This is particularly striking in an environment that was previously reserved for humans: verbal communication with customers. In fact, communication based on voice-controlled technologies, such as those we are familiar with from Siri, Alexa or Google Assistant, has recently emerged as an exciting new field of activity for banks, insurers and companies. It is in the context of developing such services that the recently introduced VA-Q testing tool is proving helpful when it comes to integrating and deploying such voice-based interaction solutions. With the new solution, companies can check all the details of their voice-controlled digital assistants. The analysis of each command in all languages is done without human intervention. This automation of quality assurance processes helps companies develop voice-driven products and services that meet customer expectations and reduce quality assurance and testing costs.

Because unlike communication between machines, there are many opportunities for misunderstanding in human-machine communication."

Due to the many variants of human action, very intelligent approaches are needed to filter out the right data – precisely the data on the basis of which decisions can be made that generate added value for the customer. Artificial intelligence can, for example, also recognize how confident the human customer is in his assessment on the basis of hesitant or confidently presented answers.

Innovations through FinTechs and digital labs

This is where the wheat is separated from the chaff: while there are currently a large number of pilot projects in the area of rule-based software robotics at major banks and insurance companies, the opportunities and risks of artificial intelligence are still being evaluated. Companies are still a long way from reliable regular operation. Much of the development work is taking place in the banks and insurance companies themselves. Various banks have established their own digital labs and departments for digital transformation in the past. However, FinTech startups often play a role as external "tenders" or "tenders" belonging to the company.

FinTechs can enrich the ecosystem of banks and – for example in an advisory context – offer financial investments for customers based on rules and insights about the customer history and customer journey."

How experimental banks are in this regard is, of course, directly related to the application: For non-binding, non-customized information, only convenience and the customer's user experience matter; for BaFin-compliant information and legally binding contracts, on the other hand, compliance guidelines in the banking environment and the security situation also play a decisive role. Because especially when external partners are added, immense challenges arise in terms of data security, privacy and compliance. And in general, of course, the liability issue of banks and insurance companies plays an important role in decisions about financial services and related advice. That's probably one reason why the financial industry has such a hard time with anything beyond purely rule-based models. 

The post The rocky road to artificial intelligence: how software robots and ki are making banking more efficient appeared first on education.telefony-taksi.ru.