UiPath Document Understanding Reviews

We use Document Understanding to validate invoices during audits. The software reads the invoices in the first step of the process, and we process them before making the payment. You must review the process and validate what's in the invoice that goes into the Oracle ERP system. We cross-check to see if the value matches or not. 

We use Microsoft Forms to capture and extract the information, so we are executing lots of documents to Microsoft Form and auditing what's in the system, comparing that to the information in the invoices.

Document Understanding processes handwritten and printed documents in PDF, PNG, and Excel format, with a monthly volume of around 9,000. We have an 80 percent success rate of automated processing.  Only about 20 percent of documents require human validation. It takes about 30 minutes per document to validate by hand. 

Before automating, we had 14 people completing the process from end to end 8 hours a day. The bots reduce that to 2 people. Our success rate is 90 percent. On some days we've had a nearly 99 percent success rate. We're satisfied with the quality.

We use the solution for purchase accounting, where we need a lot of invoices from various vendors.

The solution allows us to continue with vendors whose information comes in correctly and to stop the automation for vendors with many items that are not clearly defined. The precision is not very high when there is a lot of text in the item table.

Our clients use UiPath Document Understanding for their purchase order creations.

We need to process invoices received from vendors. This involves posting the data to SAP and creating a virtual file. To extract data from the vendor's PDF documents, we utilize UiPath Document Understanding.

The documents we process using UiPath Document Understanding are invoices and purchase orders.

The documents are in PDF format. Some documents include handwriting and screenshots.

Around 80 percent of the documents processed are completely automated without any human intervention.

UiPath Document Understanding helps handle signatures.

The call center teams automated a process where they used to manually identify configuration items in service notifications submitted by users. This manual process required a team of more than three people to analyze over 70,000 records per month. To address this inefficiency, we implemented Forms AI to automate the process. This automation has directly benefited end users.

UiPath Document Understanding has streamlined invoice processing. Previously, processing invoices was a time-consuming manual process. Employees had to read each invoice, create corresponding entries in SAP and CRM systems, and then route them to accounts payable. This required multiple resources. UiPath Document Understanding automates these tasks, reducing processing time and errors.

In the past, a team of more than 10 people was required to manually process purchase orders. Now, thanks to UiPath Document Understanding, only a few people are needed to validate the complete information and resolve any issues.

Before UiPath Document Understanding, we used over eight resources to process documents. Each resource could only handle around 20 documents per day, limiting our total daily capacity to 160 documents. However, since implementing automation, we can now process over 600 invoices daily.

UiPath Document Understanding helps reduce human error by over 90 percent.

UiPath Document Understanding has freed up staff time to work on other projects.

Our clients are satisfied with the time to value.

We use UiPath Document Understanding to extract insurance document data related to vehicle accidents in a pilot project to better understand who is at fault. We open the documents, connect with the insurance companies, and obtain the police reports describing the scene of the accident. We extract the information into a single spreadsheet to charge the third party.

Before implementing UiPath Document Understanding, humans required around 15 minutes to extract the required information per document. Now, using Document Understanding, we can process 11 documents per minute with 98 percent accuracy, saving a significant amount of time and money.

The data we extract using Document Understanding relates to vehicle accidents. The insurance company uses this data to charge third parties involved and recover most of their money with minimal human intervention.

We are extracting information from 50 insurance documents to prove the concept of showing coverage for those involved in the accident.

The information from the insurance company is in PDF format and is structured so that we can easily extract information from it. The other public policy documents vary in quality depending on the state and district in Brazil. We have 27 states, and these documents are unstructured, making it difficult to extract data from them. Sometimes they are in image format, and other times they are in PDF format.

The 50 structured PDF documents from the insurance company are all 100 percent automated using UiPath Document Understanding.

We are impressed with the machine learning and AI capabilities of Document Understanding. We can use APIs to identify whenever a vehicle accident is reported, which helps our business.

Using the AI we can access vehicle accident reports and train the machine learning to understand the majority of the documents we extract information from to help improve our business processes.

We use the UiPath platform orchestrator with an API to integrate with OpenAI.

UiPath Document Understanding has helped our organization demonstrate to potential customers how the solution, together with AI and machine learning, can benefit their businesses.

Before Document Understanding our process took 15 minutes per document and now it takes under one minute.

Document Understanding has helped free up our staff's time to work on other projects.

A recent use case was for an insurance company based in the United States. For that, we were recording or collecting the data from the insurance brokers who used to fill their documents. We had to find a few segments on the basis of them. We were collecting the data and confirming whether those brokers were coming from an authentic source. They had a stamp or a legal insurance number, and we were maintaining a few dictionaries containing the images of their signatures. Once we received a document from a broker, we passed the whole document into different segments, and then we just validated the signature part to see if it was coming from an authentic source. We validated that the signature and the image looked similar, and there was at least 80% similarity.

We were extracting the IPIN number from the Microsoft Intelligent OCR. We were able to extract almost 85% to 90% of the numbers. It contained digits that were being imposed on a stamp that we had provided to them, so there was less complexity because there was less human intervention. They were not manually writing those numbers where it could be a bit difficult for us to diagnose whether it was a four or a nine. With a digitized number imposed on the stamp, it was a bit easier for us to read it out. This is the use case that we just finished and deployed, and it is processing 150 to 230 requests on a daily basis.

I have mostly been automating banking, financial services, and insurance (BFSI) processes.

With Document Understanding, we have been able to process both structured and unstructured documents. It does not matter whether a document is structured or unstructured. The only thing is that data should be concise, and it should be constant. If we are getting 70% unstructured data and 30% structured data, we are good to go, but we should be aware of how much structured and unstructured data we are getting. If we get a picture, then based on that, we serialize them. It is either a standardized process, or we have to use some APIs or some logic to make it structured. We initially filter out based on the picture view. If the visibility of the data is less than 45% or 65%, it means that the data is not as structured. We then move it to a different folder to process it later. If it is standard and structured, we process it immediately. We do not need to worry about the chunks. There is a positive output in our hands when we have achieved 45% or 65% of our target. We can then work on the remaining part to make it more centralized, so it is a bit easier for us.

With Document Understanding, we are able to handle things like varying document formats, handwriting, and signatures. The approach we take depends on the nature of the data that we are getting. For example, a requirement from the insurance company was to mandatorily verify whether the source is authentic or not. They had metrics at their end to say who were the legal brokers and who were not legal brokers. It was not challenging for us there to extract that data from their backend because they already had all the information. We just used their APIs. We just read the data out and compared the data from there.

In terms of human validation required for Document Understanding output, we needed to finalize if the data coming from Document Understanding was correct or not. If it was not correct, we moved it to the process folder. As we marked it as incorrect, it asked us the exact location that we were looking for to get, for example, the grand total. We defined that, and then it got stored in its knowledge base system, and then it got processed. It can be processed as an attended bot or as an unattended bot. It totally depends on how much data or knowledge it has been gaining from humans, and day by day, with more knowledge, it becomes more capable of processing the data independently.

The average handle time depends on the number of cores that the operating system has. If you have 14 to 16 cores CPU in your machine, 3 minutes would be required to process a 3 MB file. It also depends on the number of pages or the complexity. If data visibility is clear and the page number is not more than five, it can process the file in 3 minutes.

After automating the process with Document Understanding, it takes two minutes to process a single PDF. I do not have the exact data of how much time humans used to take. They were probably putting in nine hours per day, and after automating the process with Document Understanding, they are putting in two hours per day, so they are saving seven hours per day. Monthly, there is a saving of 150 hours.

In terms of error reduction, in the beginning, we were getting a lot of machine errors, but as the process got smoother and the knowledge base system stabilized, the machine errors reduced, and the human errors also reduced.

Document Understanding helped free up the client's staff’s time for other projects. Before automation, they had seven people on their team, and after automating the process, they cut their budget and reduced the manpower from seven to four. They were able to free three staff members for other projects. They saved 35% to 45% of manpower.

We use UiPath Document Understanding to comb through our documents, help us prepare reports, analyze the information, and later combine the information obtained into fact sheets, results, and reports that can be used for online planning and decision-making.

We implemented UiPath Document Understanding to build an accurate and intelligent platform. We needed a platform with the necessary tools to help us automate the whole process and reduce the errors involved in document Understanding, which arise from the manual processing of our documents.

We use UiPath Document Understanding for project management, which has also been implemented in customer service and production. So it's helping us in several business processes.

We have it deployed on the cloud and on-premise. We use the cloud for remote work, and it is used on-premises at the company workstation.

UiPath Document Understanding significantly streamlines our document-intensive processes through high accuracy, ensuring reliable and well-structured extracted information. It efficiently handles a wide range of documents, including both structured and unstructured PDFs.

UiPath Document Understanding helps us process structured and unstructured PDFs as well as handwritten, typed, and scanned documents.

We process 90 percent of our documents entirely through UiPath Document Understanding without requiring human intervention.

UiPath Document Understanding has significantly reduced the need for human involvement in document processing, allowing us to focus human resources on quality control. This automation has led to a time savings of approximately 60 percent.

UiPath Document Understanding does a good job handling various document formats, including handwritten and signature formats.

Machine learning capabilities are crucial and profitable because they significantly reduce the time spent on document data tasks and the number of errors encountered. As a result, machine learning is enhancing our productivity and enabling us to adapt to change more effectively.

Forms AI has proven reliable in predicting the content of packages within our documents, significantly reducing the number of errors in our product.

UiPath Document Understanding seamlessly integrates with various third-party platforms, allowing us to combine its capabilities with web analytics tools for enhanced data analysis and processing. Additionally, its easy integration with our CRM platform further optimizes our workflow.

UiPath Document Understanding has significantly benefited our organization by enabling remote work, reducing errors in document processing, and facilitating automation. We've eliminated the need for a large workforce dedicated to manual data entry, allowing us to operate efficiently with a smaller team. Additionally, the accurate data extracted through this technology empowers other departments to work seamlessly, transforming our organization into a well-oiled machine.

Initially, in an automation process, we have human validation every 48 hours, but now that UiPath Document Understanding has been so reliable, we can move to human validation every 72 hours.

The human validation takes three to four hours to complete.

UiPath Document Understanding has significantly reduced human error due to its accuracy.

Before implementing UiPath Document Understanding, processing ten documents took eight hours. Now, with UiPath Document Understanding, the same ten documents can be processed in less than thirty minutes.

The most beneficial aspects of UiPath Document Understanding for our data extraction needs have been the UiPath Academy courses and the free trial, which proved to be an eye-opener.

Machine learning has significantly benefited our organization by enabling the automation of numerous processes. We have successfully implemented machine learning to automate not only document generation but also report writing, analytics, and production. Furthermore, our CRM and planning processes have been streamlined through automation, which is made possible by machine learning. This technology has proven to be invaluable to our operations.

AI is helping us power robots and even expand our system capacity. It is the driving force behind remote working and also the driving force behind minimal human validation.

In a medical healthcare department, when we need to retrieve digital documents, we need to classify them. The first step is to use AI to understand what type of documents we're dealing with. Once we've identified the template, we can extract information using specific OCR tools. Depending on the confidence of the extracted results, we may need to apply additional OCR, use a more active tool, or pass the document to an agent for review if the AI doesn't recognize a specific element like the "person page of the commission." Finally, the extracted fields are classified within the system and organized into different folders. This is the process I'm using with UiPath Document Understanding.

Document Understanding can complete each document within one second.

It can be applied to the healthcare industry to streamline the processing of medical documents. This includes scanning and applying OCR to convert physical documents into digital formats.

We can tune the AI component to improve the quality and accuracy of the documents being processed.

Typically, the AI process involves several steps. Firstly, it recognizes the template, which essentially identifies the input format being used. Secondly, it applies rules configured in a JSON file. This file specifies details like the expected fields for the recognized template, such as name, age, date of birth, and security address. The AI then reads and analyzes data from the specified location based on the recognized template. It applies the predefined rules to extract relevant information and search for the required fields. If the input doesn't match any known template, it employs more general search methods to locate the desired information. This is the core functionality of the internal AI component.

Of the 1,000 documents we process, 90 percent are completely automated.

My three OCR tools each incorporate three AI components. These components work in tandem, with the activity determining which AI component takes the lead. For example, if the first AI requires a minimum accuracy of 86 percent and encounters text with 85 percent accuracy, it passes the task to the next AI component. This next component employs a different OCR tool in an attempt to achieve the required accuracy. If it still falls short, the task is then routed to a human agent.

Our integrations leverage robust API connection services. A single, secure authentication method protects access to JSON files. Requests are sent and product responses are seamlessly handled. This API-based approach provides faster and more efficient integration compared to manual interface interactions.

UiPath now includes a document understanding AI components, eliminating the need for third-party solutions like ABBYY. This allows for quick and automated extraction, analysis, and template recognition of information from various documents. By training the system with diverse examples, the AI component can become highly efficient, similar to ABBYY's global OCR capabilities. This is a significant improvement, as it eliminates the need for additional integrations like ABBYY within UiPath projects.

I found UiPath Document Understandings' ability to understand varying document formats to be good. I had no issues with the templates I was using.

Using AI and machine learning can significantly speed up the recognition of new formats, templates, customers, or entities introduced into our process. It is particularly beneficial when dealing with low-quality documents, which often require manual intervention. By implementing a machine learning model at the beginning of the process, the system can learn from successful agent solutions and incorporate them into future scenarios. Clear feedback, including agent ID and task details, further enhances this learning process. As a result, machine learning can help save time, reduce costs, and improve overall process accuracy. This makes it a valuable tool within UiPath.

Less than ten percent of processed documents require human validation. However, when customers provide input that falls outside pre-defined templates the usual 90 percent of cases, the system cannot recognize it and fails to notify agents. This means a new template will be implemented to include human-agent collaboration when training AI models.

The validation process depends on the specific template and the data being acquired. If all data is extracted from the entire template, the validation process can take less than one minute.

The manual document process took us around ten minutes and now with UiPath Document Understanding, the process is within seconds.

Since implementation, human error has been reduced by 30%.

UiPath Document Understanding has helped save 50% of our time in instances when no human validation is required.  

We primarily use UiPath Document Understanding for finance processes, covering both transactional procedures and reviews. One recent example involved streamlining the onboarding process, including pre-boarding, onboarding itself, and post-onboarding follow-up. The company typically requests various documents from applicants, which are then processed manually. However, due to variations in country-specific standards and requirements, HR personnel often spend significant time handling these documents.

Our solution involves creating a seamless online portal where applicants can upload their documents. These documents are automatically screened by the system and directly uploaded into the company's EFP system. This significantly reduces manual work for HR and finance teams. Similar automation applies to processing invoices from various suppliers in different formats. We leverage machine learning tools to train the system to read documents with varying complexity levels.

Essentially, the system mimics how an HR professional would process documents, capturing their knowledge and integrating it into the automated workflow. This reduces processing time and workload for both the company and its clients. Our focus lies on automating tasks within well-defined contexts, making us less involved in product development activities at this stage.

Initially, our clients were primarily interested in UiPath Document Understanding out of curiosity about its potential. Their main focus was on automation, but we also engaged in discussions about the broader benefits, such as time savings. We highlighted that a 30 percent time reduction allows them to focus on tasks with higher value. However, what I found even more crucial was the impact on lead times. Manual processes often lead to work stoppages, delays, and roadblocks. Automation, even partial, can significantly reduce lead times. For example, a task that previously took five weeks can now be completed in just a few days. While security concerns may necessitate some manual intervention, such as allowing the head of HR to retain some oversight, the overall process becomes more streamlined over time.

Most document processing is automated, improving efficiency and ease, especially in back-office transactions. However, areas like marketing, where business plans require creativity and flexibility, remain manual for now. Where documents are stored, and manipulated, and data needs to be extracted and distributed across various systems, the process is often cumbersome. Traditionally, someone would manually open each system, which is time-consuming, especially considering most companies have hundreds of them. This is where tools and systems come in, able to connect across platforms, read data from various sources, and make interpretations. The level of automation depends on the company's maturity. Sometimes we leverage their existing data, while other times we implement techniques to extract more insights. Ideally, we'd be able to predict and anticipate future needs, but for now, with clients, we're primarily focused on analyzing data and helping them automate their processes. This is the first step.

The volume and types of documents we process with UiPath Document Understanding vary depending on the client. For smaller companies with a few hundred employees, the needs are different than for large international corporations with thousands. These international clients often have diverse locations with varying processes and systems, making automation more challenging. In HR departments, for example, the sheer number of applicants and their associated documents can be immense. Ensuring accuracy is crucial, as mistakes can have significant consequences. Finance departments also present unique challenges, as data might be hidden or incomplete. This requires them to be at a certain level of maturity to benefit from automation effectively. The complexity of documents is another key factor. While machine learning can handle many documents, it has limitations. Some documents might be too time-consuming to train on, making the investment in automation impractical. This can leave a portion of documents requiring manual processing. Overall, UiPath Document Understanding automates the processing of the majority of documents we handle, around 80 percent. However, for the remaining 20 percent, manual intervention is still necessary due to document complexity, data limitations, or training time constraints.

UiPath Document Understanding helps us extract data from various document formats, including tables, handwritten content, checkboxes, and barcodes. However, poorly legible documents present a challenge. Automating 100 percent of documents is currently impossible due to diverse languages and handwritten sections. Our current approach categorizes documents into easy, medium, and complex based on difficulty. We prioritize easy documents as complex ones require significant time investment with uncertain results. Unfortunately, machine learning for document processing can be time-consuming. We prioritize documents based on return on investment. For example, if we have 10,000 documents, we might skip two unique ones, even if theoretically similar to others. If only two or three data points are needed, but the structure drastically varies, processing might not be worthwhile. Imagine a 10-page phone bill invoice with a minimal value of €10. Investing time in such documents offers a minimal return. Therefore, we focus on documents offering greater value.

Around 70 percent of the documents are processed automatically using UiPath Document Understanding.

UiPath excels at connecting with various systems compared to some competitors. This is crucial when promoting it to clients, as in our case with our UiPath partnership. All our developers have UiPath training, and we strongly believe in its capabilities. However, internal legacy systems within companies can pose challenges. For example, a client with an EFP system they plan to replace might hesitate to automate now. Integrating UiPath with basic IT infrastructure is essential, and frequent system changes demand flexible solutions. While UiPath is adaptable, we need to demonstrate its compatibility with various systems to gain client buy-in. This will make them more open to automation. It's important to remember that company maturity levels influence their automation openness. While UiPath has no control over that, adapting to ever-changing environments requires flexible systems. By showcasing UiPath's ability to work with different systems, we can overcome client hesitation and secure their trust in our proposed automation solutions.

It typically takes clients about a month to see the benefits of UiPath Document Understanding. We start by showing a demo. We often use the UiPath website itself for inspiration, and we also consult with UiPath staff to see if they have any pre-built demos for specific areas, such as onboarding. We create short, simple videos tailored to their needs and showcase them to both HR and IT personnel, giving them a glimpse of the solution before implementation. While deployment ultimately requires its timeline, we can typically craft a process description within a couple of weeks, allowing for a swift rollout. The tools themselves are relatively quick to use. In my experience, the main bottleneck usually lies within the client organization itself. Functional teams are often busy, have competing priorities, and sometimes change their decisions. Navigating these internal dynamics can be time-consuming. The actual development time for tasks like process mapping, decision-making, and technical implementation is relatively short, typically measured between 10 to 20 days. However, building consensus, convincing stakeholders, and developing a compelling business case can take considerably longer. Internally, clients often encounter both promoters and detractors – individuals who welcome or resist change. These internal dynamics are often the biggest hurdle. However, once the decision is made, we can quickly create a targeted demo showcasing the added value UiPath Document Understanding can bring.

On average, human validation takes just a few minutes. Additionally, the number of full-time equivalents was reduced by 30 percent - that's a significant achievement. Lead time has also decreased dramatically, much more than the FTE reduction. A small department of three people can now do the same work with two, freeing up one person for other tasks. It's important to note that lead time reduction depends on the specific case. Theoretically, in a perfect scenario with seamless workflow, automation tools operating 24/7, and no disruptions, a five-fold decrease in lead time is possible. However, real-world scenarios often involve unforeseen issues requiring manual intervention, limiting the maximum achievable reduction. Still, significant lead time reductions are attainable through consistent improvement efforts.

When it is done well we can reduce and improve the accuracy through automation helping to reduce human error.