Use cases

How customers are leveraging Applica’s technology to optimize their automation

Situation

An employee at a large financial institution needs to run an assessment in order to determine whether or not to give a bank guarantee. This process requires manual scrutiny of complex corporate documentation and correctly dispatched documents to different categories of bank guarantees.

Problem

The manual process is expensive as is executed by a qualified risk assessment employee. Manual extraction of information is also slow, error prone, and requires significant human effort.

Solution

Applica RTA categorizes all of the bank guarantees into six independent baskets. Each basket represents a different type of bank guarantee (e.g.: Performance guarantee, Bid bond guarantee, Financial guarantee, Advance payment guarantee, Foreign bank guarantee, Deferred payment guarantee). Each category requires a unique type of documentation, which Applica RTA will automatically categorize into the correct basket. Furthermore, Applica RTA will create a subcategory of each basket and extract the required information from the different categories to run a full risk assessment.

Result

The turnover time to process risk assessments changed dramatically with the implementation of Applica RTA. With all the extracted data from a range of documents located in a centralized repository, it is easy for employees to query the required information on-demand and determine whether or not to give the bank guarantee within seconds.

Situation

A financial intuition needed to review non-disclosure clauses in 60,000 US client accounts. Previously the organization employed a team of 12 attorneys to review the documents manually and create an Excel spreadsheet with the results.

Problem

The organization had a massive repository of poorly scanned client documents and no one knew what was inside of it. When an executive asked the team what their obligations were in terms of non-disclosure on these accounts, it took the team of 12 attorneys six weeks to determine an answer.

Solution

Applica RTA took the historical non-disclosure agreements, extracted the required information, and put it all in a structured, searchable form. Now whenever anyone has a question on these documents, they can execute a simple query and have an answer the same day. New non-disclosure documents are also put into this structured database for self-service use.

Result

By automating this process and making it easy to access for repeatable future use, the organization no longer needed to employ the team of 12 attorneys to review the data, reducing costs by 80% and turnaround time decreased from six weeks to less than one day.

Situation

A telecommunications company received 400,000 customer requests per month. The customer care staff were spending 60% of their time reading/sorting inbound messages, as there were often duplicate or triplicate claims for the same issue in the system, a consequence of customers not receiving a response fast enough and submitting through multiple channels.

Problem

The department was operating at full capacity but turnaround time to manage claims was slow and often exceeding the stated time they gave to customers. The company employed 250 full-time staff members, but still wasn’t able to scale appropriately to meet influxes in claims – such as during a service disruption.

Solution

Applica RTA was deployed to divert 60% of the inbound requests away from the customer care representatives and also set up automated messages to update customers if the response time would be longer than normal. Applica’s unparalleled AI was able to fully automate replies for 35% of the messages, remove the 10% of duplicate messages, and the remaining 55% of messages that required staff member review were partially automated for more rapid processing. E.g.: Applica RTA extracts and classifies the relevant content so that the agent can respond 30% faster to the claim, as they no longer had to read through the entire request to understand the issue.

Result

The implementation Applica RTA resulted in $5.5 million in cost savings for the 250-person customer care department and led to a complete transformation with how representatives spent their time, greatly increasing customer satisfaction due to much quicker response times. Additionally, Applica RTA helped the company identify the most pertinent requests requiring expeditious handling, avoiding potentially costly legal and/or regulatory issues.

Situation

Insurance and reinsurance companies often receive applications and supporting documents from clients/policyholders, agents, and brokers that are paper based, with data that must be rekeyed from the source documents for operational processing. Data may be structured, semi structured (e.g.: loss reports, insurance contracts) or unstructured (e.g.: emails, property inspections), provided in a PDF format (either “selectable” or scanned image PDF) and/or Excel. For underwriting, data for a rated entity may need to be extracted, identified, and aggregated from multiple forms or sources for the same submission.

Problem

The manual process is expensive as is executed by a qualified risk assessment employee. Manual extraction of information is also slow, error prone, and requires significant human effort.

Solution

Applica RTA categorizes all of the insurance forms/documents into independent baskets. Each basket represents a different type of insurance group (buildings and vehicles and companies, etc.). Each category requires a unique type of documentation, which Applica RTA will automatically categorize into the correct basket. Furthermore, Applica RTA will create a subcategory of each basket and extract the required information from the different categories and passed extracted information to domain systems (CRM, risk assessment, etc.).

Result

With the implementation of Applica RTA, all data from various insurance forms were structured into a standard, searchable format. This not only decreased errors innate with human work, but also decreased the time required to QA the documents and sped up the end to end processing. Employees that were once stuck keying in information to systems were redeployed to more valuable tasks.

Situation

A multinational organization processes 45 million invoices per year for their clients and their current automation solution is a template-based OCR software that achieves approximately 65% accuracy when it comes to data extraction. As unstructured information is consistently contained on invoices, humans must be involved to deal with the balance of cases before they can be put into the payment system – not allowing for full automation, optimal efficiency or lowest cost to process.

Problem

OCR template-based solutions have a lower level of accuracy and precision when it comes to extraction of the key values from frequently varying invoice types. They also aren’t able to comprehend the unstructured data that may be essential in reconciling the invoice (e.g.: a legal clause, annotation, etc. added to the bottom of the page). OCR template-based solutions, even when applying ML and AI, can’t interpret text or make key decisions regarding validation, booking, and approval when applying semi-structured and unstructured data.

Solution

Applica RTA is fully templateless allowing for the handling of any structure or format, including complex tables. Applica RTA leverages proprietary 2D Contextual Awareness and layout aware Language Modeling (LAMBERT) to review invoices and make decisions similar to a human mind. No longer needing to adhere to the rigid structures or rules of template-based technology, Applica RTA produces better results than software specifically designed for invoice processing.

Result

By being able to process all data types, the level of precision on extraction increased from 65% to over 95%, this dramatic increase in precision on data allowed for straight-through processing and the highest levels of efficiency and automation ever attained. Therefore, the cost per invoice decreased significantly, as necessary human review went from every 4 in 10 documents to less than 5 in 100 documents.

Situation

A multi-national bank runs a KYC process that requires manual scrutiny of 1,200 complex documents per month.

Problem

The manual process is expensive as is executed by qualified lawyers. It is also time consuming as each document takes 48 hours to process, and fallible as it relies on numerous external factors to work properly.

Solution

Applica RTA was trained with an extremely small number of examples and achieved a 1:1 replacement of human lawyers.

Result

Automation accuracy levels are 98% and the platform learned from the 2% of corner cases for future use. The document turnover time decreased from two days to just five minutes, greatly increasing the bank’s customer service level.

Situation

One of the largest Debt Collection Agencies in Europe manually processes more than 200,000 unique legal documents per month.

Problem

Debt collectors typically purchase debt portfolios in batches and the workloads may vary significantly. This is hard to plan for so the agency must remain fully staffed at all times. Additionally, within debt portfolios certain cases must be prioritized to meet regulatory deadlines, while others can be handled within due course. Human errors are particularly costly, which lead to loss of revenues and delays.

Solution

Applica RTA was deployed in an AWS secure cloud. The solution is being used via standard APIs accessed by the customer’s document management system. Maintenance of the solution is performed through interactions with business users running quality control (results of the quality control are fed back to the platform for continuous improvement of the AI models).

Result

The agency automated 90% of what was previously done manually and eliminated 85% of human errors. The document turnover time decreased to less than one second per page and 90 full time employees were reassigned to higher value work.

Situation

Many of the world’s largest economies only have until the end of 2021 to implement LIBOR Fallback Contract Language in a wide array of their financial instruments, including commercial loans, syndication, bonds/floating rate debt, securitized products, mortgages, derivatives, and interest rates/swaps.

Problem

Existing contracts must be amended to reference rates other than LIBOR. To do this with human workers would be nearly impossible in the time allotted for full implementation of the new rates. An organization’s existing RPA solution cannot process these documents in a fully automated manner and it would still require significant human intervention and manual effort.

Solution

After an RPA bot retrieving required documents from an archive is implemented, Applica RTA interprets and classifies the reference rates as required. This allows for minimal human interaction throughout the entire process – regardless of file type, language or layout.

Result

Streamlined the process of contracts amendment, which reduced the involvement of qualified counsels by 75%. The faster document turnover time (less than one second per page) enabled the processing of tens of thousands of documents per day.

Situation

A regional bank runs a mortgage lending process that requires a specialist to review large volumes of paper documentation in order to extract information from property sales contracts.

Problem

Manual extraction of information is slow, error prone, and requires significant human effort. The resulting lengthy turnover time has a negative impact on the customer experience. Existing automation options require too much engineering effort for deployment and maintenance at a small organization.

Solution

Applica RTA was deployed in the cloud and trained to extract required information from property sales contracts, regardless of the layout and wording.

Result

Deployment of Applica RTA saved 80% of the human effort related to extraction of information from the property sales contracts and reduced the time required for mortgage lending decisions by 30%.

AI-based text comprehension

Connect with an Applica expert to discuss how easy it is to automate your text-intensive business processes

Get a demo