Intelligent Process Automation (IPA) is technology that enables organizations to automate processes that involve structured, semi-structured, and unstructured document formats including forms, documents, text, images, video and more.
Among the list of intelligent process automation topics covered on this page:
Intuitive interface enables process owners to build their own automation models; no data scientists required. Automate workflows involving unstructured data to dramatically reduce processing time with unmatched accuracy.
Automate the process of converting unstructured data to a structured format suitable for analytics engines – gain new, actionable insights.
Award-winning AI technology can “read” unstructured documents – so your employees don’t have to. Build models with just 200 sample documents – no rules or templates required.
Now that many organizations have reached the limit to what RPA can do for them, they are asking – what is Intelligent Process Automation? The main purpose of Intelligent Process Automation (IPA) is to enable organizations to automate processes that involve unstructured data, including text and images. It does so without requiring rule-based decision-making or huge training data sets that are out of reach for 95% of enterprises.
Indico’s approach to process automation builds on the artificial intelligence concept of transfer learning, where a model trained on one task is used for another, related task.
Transfer learning addresses one of the key challenges in any AI solution: the time required to learn exceptions. An IPA solution would normally have to understand thousands of use cases before it could be used in production to automate an actual process. Transfer learning changes that equation.
Indico created a base model consisting of more than 500 million labeled data points, enough to enable the model to understand human language and context. Applying transfer learning enables users to then create custom models for downstream tasks using a fraction of the data normally required – 100x to 1000x less as compared to traditional approaches.
Rather than training the model on hundreds of thousands of examples, or creating rules to account for every variation of the documents at hand, Indico’s intelligent process automation tool enables you to start with its base model and train on just 200 or so examples of the process you want to automate. In just an hour or so, you’ll have a complete workflow automation model.
Also, because most of the training is already done up front, the unstructured data platform can run on just one or two GPUs, and scale up using low-cost CPU. Overall, you get a highly effective intelligent process automation tool that can likely pay for itself in short order by dramatically reducing both process cycle times and the human resources required to perform the process.
That is the benefit of Indico’s intelligent process automation tool; it doesn’t require a million-dollar investment to run.
Intelligent Process Automation offers a single solution for document intake, understanding and digitization: in other words, unstructured document processing, as well as processing of structured and semi-structured documents. This allows for the end to end automation of common document-based processes in enterprise business, including contract analysis, customer on-boarding, commercial underwriting, financial document analysis, mortgage processing, billing form reviews, insurance claims analysis and much more. With its cognitive intelligence capabilities, IPA can understand the text, images, documents and other unstructured data that are fundamental to so many business processes – and make accurate judgments based on surrounding context.
Unstructured data creates problems for rule-based automation engines, including robotic process automation platforms, and OCR templating approaches, because it’s so difficult to define rules that apply to something you can’t predict. And unstructured data is nothing if not unpredictable.
By definition, unstructured data refers to data that is variable in nature. It could be contracts, Word documents, text (including emails) and images
IPA also enables companies to analyze unstructured data. Because business intelligence and analytics tools can generally only deal with structured data, the value inherent in unstructured data remains untapped. Indico Data changes that equation by automating the process of turning unstructured data into a structured format, such as JSON or .csv. Once it’s structured, you can feed the data to analytics engines and business intelligence tools, such as Microsoft Power BI and Google’s Looker.
Along the same lines, Indico Data can prepare unstructured data for input to data visualization tools such as Tableau, which also requires data to be in a structured format. Effectively, Indico Data makes possible unstructured data analytics and visualization.
This is a sea change for companies that have been relying on rule-based automation engines, including robotic process automation platforms and OCR templating approaches. These tools cannot effectively process unstructured data because it’s simply too difficult to create rules for data that has no predetermined format.
By definition, that describes unstructured data. It is variable in nature, including virtually any kind of content, from Word documents and PDFs to emails, images and even video.
With IPA, you can now automate processes involving such content, as well as unlock the value inherent in your unstructured content and feed it to BI and analytics tools. Literally decades-worth of business intelligence is now newly at your disposal.
The process through which companies use Intelligent Process Automation to build data models is simple and highly effective. Business subject matter experts label the data points they deem most important to whatever process they’re looking to automate. As they apply labels, the model is updated on the fly and will start to show predictions on subsequent datasets. Once you’re comfortable with the predicted results, you’re done building your model.
The beauty of this approach is that the people who understand the business problem and the desired results – those on the business side of the house – are the ones who train the model. We call them “citizen data scientists.” With Indico, there’s no need to try to explain to a data scientist what you’re after and then hope you get the appropriate results. As a citizen data scientist, you can create models yourself.
And it’s not a complex process. Everything is in plain English and you can have a fully working model in an hour. Intelligent Process Automation is just that simple.
If that sounds different from other artificial intelligence solutions you’ve encountered, that’s because it is. While Indico’s IPA solution is certainly sophisticated in its use of cognitive technologies including machine learning and natural language processing, we keep the technology behind the scenes, enabling an army of citizen data scientists to use the technology to solve real business problems.
Natural language processing (NLP), for example, is core to our IPA platform. It’s what enables our generalized model to understand the context around unstructured data, just as a human would. But it’s built into our models and functions behind the scenes; there’s no need for those who use the platform to even know what NLP is.
The same goes for machine learning (ML). While our engineering team built our IPA platform using cutting edge ML models, they all sit in the background – there’s no need for users to tweak or otherwise interact with them, or even understand how they work. Citizen data scientists instead can just think about how to apply IPA to take repetition and complexity out of their processes and deliver real business benefits.
IPA can be used to automate the classification and annotation of a new claim, and route it to the appropriate SME for evaluation and processing. The result is faster turnaround time and improved accuracy in claims processing, which drives improved customer satisfaction and organizational efficiency.
One of the thorniest parts of the commercial insurance underwriting process is getting an accurate picture of the applicant’s loss history, generally gleaned from loss run reports . But it can be a cumbersome process to collect all the reports and accurately extract data from them for input into the underwriting system – making it an excellent candidate for intelligent process automation in insurance .
Few vertical industries are as document-intensive as healthcare, whether on the provider or insurance side. That makes the healthcare industry ripe for tools that can automate insurance claims processing and other chores, for both providers and insurers alike. Intelligent process automation can help healthcare organizations address unstructured documents driving cost savings and improving the patient experience.
Getting new clients is a good thing, but for insurance companies it also creates a challenge: processing all the required documents. To date, it’s been a largely manual process that for large insurers can easily involve 15 million documents per year, making it a ripe target for intelligent document processing technology
Major commercial underwriting processes often involve thousands of pages of documentation. Insurance workflow automation can dramatically improve the process by creating underwriting criteria that IPA solutions automatically recognize, enabling them to quickly come up with a “score” for each potential customer. The result is a major reduction in response times to customers as well as improved accuracy, satisfaction, organizational efficiency and profit.
IPA can be used to automatically classify and extract relevant unstructured data from customer onboarding documents into the bank’s digital management system. This results in improved accuracy and speed for onboarding a new customer, driving improved customer satisfaction and faster time to revenue for the bank.
In 2017 a UK-based regulatory group announced the LIBOR interest rate benchmark would be phased. As a result, banks and financial institutions around the globe are scrambling to determine their exposure – a task that’s tailor-made for intelligent process automation solutions.
Banks with detailed processes for appraising and approving mortgages, including data extraction and image recognition, can use intelligent process automation tools to automate the process of extracting relevant unstructured data from onboarding documents, as well as to analyze images. IPA can be used to bring workflow automation to the mortgage approval process, allowing it to become far more efficient and consistent
A lockbox is a service that financial institutions including commercial banks offer. Similar to a post office box to which companies have customers send correspondence, a lockbox is a service offered by commercial banks whereby companies can have their customers send payments to the bank. For a fee, the bank takes care of matching each payment to an invoice, helping to streamline the accounts receivable process for its client company.
When it comes to automating document processing, titles and deeds in particular present a vexing challenge. The reason is simple: these documents vary enormously depending on where they come from. Naturally, each county has its own forms for titles and deeds, and they are not all alike – far from it. Businesses that need to process lots of these kinds of documents have historically had to manually extract data from these forms and enter it into spreadsheets and other financial systems. Clearly these firms could benefit from process automation.
Investment firms can use IPA to analyze the financial health of companies before deciding whether to invest in them. Instead of poring over thousands of financial statements and manually extracting relevant data from each of them, IPA enables financial firms to automate the process, pulling out relevant data and normalizing it for insertion into data processing tools. The result is a dramatic improvement in speed and efficiency.
Investment firms receiving trade processing documentation via email and PDF formats can use artificial intelligence automation tools to extract relevant unstructured data from these trade documents and compile it into a normalized format that can then be integrated with the firm’s digital management system. IPA enables companies to eliminate untold hours of manual data processing. Other industries where IPA is being applied include legal services and marketing (CRM).
Processing invoices is an issue that just about any large company struggles with – and one that’s ripe for automation. But it’s a classic example of an application where process automation software that relies on templates fails to deliver consistent results, for reasons that are easy to understand. When you have invoices from many different companies, you’re essentially dealing with unstructured data, which rule-based tools and template-based approaches to automation aren’t well-suited to handle. Intelligent process automation (IPA) tools, on the other hand, can handle unstructured data. IPA uses OCR, machine learning (ML), and natural language processing (NLP) to enable it to understand the context in a given document, enabling it to identify the relevant information you want to extract, without having to create a template for every variation of the invoices in question.
In the corporate Inbox use case, an intelligent process automation tool would be able to “read” an incoming email, discern what the topic is, then route it to an appropriate subject matter expert. For relatively simple matters, such as a change of address request, the IPA tool could extract the pertinent information and input it into an appropriate downstream system. An intelligent process tool can also extract and automate the handling of any attachments from an email, such as PDFs and Word documents. Here again the tool is smart enough to “read” the attachments and extract relevant data for input it into another downstream tool for processing or future reference, such as a customer relationship management (CRM) system.
Extract relevant data from insurance documents and feed it to an analytics engine to identify any areas where you may be running afoul of regulatory requirements. Indico Data can also help index data to simplify and speed responses to regulatory bodies.
Intelligent Process Automation can streamline processes involving unstructured documents from receipts, purchase and sale agreements, images, and contractor estimates. Reduce processing time by up to 85%.
Extracting data from rent rolls and converting to a structured format enables you to feed it to an analytics tool to gain insights into cash flow, turnover rates, vacancies, and opportunities as well as speed due diligence checks.
ISDA Master Agreements, which spell out the terms between two parties regarding over-the-counter (OTC) derivatives transactions, are notoriously complex and lengthy. The latest version, from 2002, is 28 pages long. Yet, it’s common that multiple subject matter experts (SMEs) must examine the documents and confirm terms are correct before executing a trade. Indico Data automates the process, dramatically reducing ISDA Master Agreement processing times.
Intelligent Process Automation can transform the verification and processing of client documents and monitoring of negative information. Indico Data machine learning models can help identify the warning signs of a money laundering risk by comparing to baseline data. You can also extract key data from various documents and feed them to an analytics engine that can identify suspicious activity.
Intelligent process automation builds upon existing processing automation technologies that also sought to streamline business processes, namely business process management (BPM), business process automation and robotic process automation (RPA).
BPM and business process automation is focused on improving an existing business process. That often involves automating some steps in the process, although that’s not necessarily a requirement. It’s more about optimizing a process to make it more effective and efficient, often by using methodologies such as Six Sigma and Lean.
As its name implies, RPA does involve process automation and works well with repetitive, deterministic business processes involving structured data – where there is no judgment involved. Tell it exactly what you need it to do and RPA can do it better, faster and cheaper than a human.
If a task comes along that deviates from the pre-defined task, RPA will not be able to automate it. It cannot make judgments about information or learn and improve with experience. In that sense, RPA is different from machine learning, and IPA.
For the same reason, RPA is ineffective with workflows involving unstructured data – those that require some level of cognitive ability. And this type of data makes up over 80% of the data in most enterprises today.
Because of IPA’s cognitive ability, it is very well-suited to work with business processes involving unstructured data – all the text, documents, and images that drive many enterprise business processes today.
What is intelligent process automation?
IPA does not replace or compete with RPA. It complements it, handling the unstructured data, with the output being structured data that can be re-inserted back into a business process that RPA can then address, leading to true digital transformation. IPA can pick up where RPA hits a roadblock in such diverse use cases as customer communications, report aggregation and insurance claims.
We realize those are some rather impressive numbers in terms of return-on-investment – but they are most certainly real.
Suppose a given process involves 10 employees who each make $100,000/year, or $1 million total. The team performs 500,000 tasks per year dedicated to this process, so the cost per task is $2. Let’s say an IPA solution can automate 75% of those tasks, which is not at all unrealistic. The cost per task falls to just 50 cents and your annual gross savings is $750,000. Subtract the cost of the automation solution and you can calculate your ROI. (Hint: it will be huge.)
At the same time, you’re gaining soft benefits including increased employee satisfaction and productivity – because employees won’t be doing the same monotonous tasks every day, instead taking on more rewarding work. In the example above, you now have $750,000-worth of employee time to dedicate to other areas, dramatically increasing the capacity of the organization.
What’s more, the newly automated tasks will be performed with increased accuracy and consistency, which likewise saves money and helps ensure compliance with industry regulations.
Don’t take our word for it. We recently sat down with MetLife’s VP of Intelligent Automation to discuss their automation journey from solving simplistic tasks with RPA to deploying Intelligent Process Automation to automate unstructured document-based workflows. MetLife has found $100M in value through hours saved that they can unlock in the next 5 years for their businesses by using Intelligent Process Automation on unstructured data. You can watch the full interview here.
Indico Data essentially turns those platforms into unstructured data analytics tools, enabling them to process unstructured data. Now you can unlock the insights trapped in your collection of enterprise documents, raising the potential for new – and potentially lucrative – business opportunities.
85% reduction in process cycle times
Drive customer satisfaction and quicker time to market for new initiatives
4x increase in process capacity
Scale critical processes without increasing expenses, for more cost-efficient back office functions
80% reduction in human resources
Free up critical resources to work on higher value-add projects rather than repetitive low-value tasks
Ease of use
No data science expertise required
1000x less training data required
As compared to traditional artificial intelligence solutions
Built for unstructured data
Works with text, documents and images to automate almost any business process