Head of Communications Industry Iberia & LATAM at CELONIS INC
User
Oct 13, 2021
Let me start by saying that it's just a catchy word, like many others in our industry. Depending on who you ask (there are some good answers here already) you will get a different answer.
I would summarize it as the concept of applying different AI and other data science-related methods to enhance how you typically envision task and/or process optimization initiatives. From how you define and design your future process implementation, to how it will actually perform some of the tasks involved, or how your implementation adapts to changes over time.
So it's a super broad spectrum, where different vendors will pull the concept to their area of expertise.
I'll try to summarize some real-world examples across that spectrum:
- Automated/assisted current process discovery, benchmark and diagnostics, to accelerate your process transformation efforts, and feed that into your process execution stack.
- Process simulation / digital twin simulation, to support strategic decision making based on sophisticated business simulations (like a super "Excel What-if" feature but at company scale) to design optimal TO-BE processes.
- Automated/Assisted task flow design (based on previous topics), to ease implementation of currently manual task sequences into your task/process automation tools.
- Smart/dynamic run-time decision making, versus static, rule-based decision making (like sending a customer request to a different queue based on predictive churn risk, instead of static rules, like customer scoring.
- Process self-healing/adjustment: Use ML to let the process self adjust based on past experience to optimize outcomes. A basic example would be around decision-related thresholds that are adjusted based on past decision outcomes.
- Actually performing tasks automatically, not that easy to automate by "older" approaches. Like validating an ID with computer vision technologies (not just reading the ID, but actually validating it's legit), instead of sending those to a pool of humans for verification.
- Enriching the context data set the process can use to take optimal action, like automatically extracting customer intent from voice/chat, or applying OCR to read a scanned invoice.
- Smart monitoring & management of your process automation platform. An example would be to manage resources dynamically based on learned patterns, and avoid unnecessary resource consumption, or adjust for a process load hike based on past behavior.
I hope this helps understand not the term, but the actual potential behind it, regardless of who's trying to sell you what ;-)
Search for a product comparison in Robotic Process Automation (RPA)
Intelligent Process Automation is basically an upgraded form of robotic process automation. Unlike RPA, intelligent process automation can understand context, learn, and iterate. IPA can also handle both unstructured and structured data and supports some level of informed decision-making. Informed decision-making can further be divided into task level or process level automation. IPA helps organizations to access and analyze unstructured data like images or text that is inaccessible by other means to gain important insights.
IPA can take unstructured data and turn it into structured data for use with RPA technologies. For reasons like these, the technologies are not mutually exclusive but can work together to optimize business processes.
So, while we may be a long way off from robot servants that do our laundry and mow our lawns, robots are already and will continue to play increasingly important roles in our daily lives. Robotic process automation makes work less tedious and allows organizations to scale their operations to provide more value to customers. Intelligent process automation builds upon RPA, giving systems the ability to automate tasks and to learn from them.
For organizations that are new to automation, getting started with RPA technologies is the way to go. They are easy to use and can be implemented with low-code business process management software. Once automation has been introduced into one or more processes, organizations can consider introducing intelligent process automation (IPA) for complex business processes.
Use cases of IPA
Comparing Prices
Companies generally engage in large purchases to offer varied services or products. The cost of these purchases can affect the company's revenue or profits and hence companies tend to research online for making informed decisions. This exploration for right decision making can take a long time and involves complex processes, which is the reason why many companies are now adopting IPA. The right IPA solution can not only compare prices from different vendors but can also compare product attributes and quality. Companies employing IPA can now buy the best services or products at the best prices possible.
Customer Information Storage
IPA can help you store, organize and categorize all types of customer information to ensure easy access to everything. The system can automatically categorize different data sets such as contact information, purchase history, preferences or personal information. The system can display all information to customer support team, sales representatives and similar employees. There is no need to enter this information manually or worry about its accuracy. IPA can always be depended for accuracy in contrast to employees and has a lower error margin. IPA can thus help you reduce repetitive task saving you from unnecessary stress and labor involvement.
Recruitment Process
Recruitment is another prime candidate for implementation of IPA system in large organizations. IPA can aid the recruitment process by helping the HR team to source resumes from a range of online platforms, analyze the exact skillsets, derive value, sort through spams and unwanted applications. It helps the HR team to gain access to the right skills pertaining to candidates with much reduced effort and cost efficiency. IPA not just filters unwanted applications but also helps HR to sort through the right resumes and have access to every application exactly suited to the job requirement. IPA helps the HR team throughout the recruitment process from screening to assessment to final onboarding and administration.
I think the market largely has the wrong definition of IPA, but we can turn to HFS for a take on it: "
"Intelligent Process Automation Definition: Intelligent Process Automation (IPA) is the use of technology to allow a business function or part of the operation of a process workflow work automatically. It includes the use of RPA, BPM suites, Remote Desktop Automation, screen scraping and custom scripting and related technologies." (HFS, 2018?)
I think it is fair to conclude that
Tech Target: "IPA is designed to assist human workers by doing manual, repetitive and routine tasks that were previously performed by humans. Technologies combined in IPA include robotic process automation (RPA), artificial intelligence (AI), machine learning and digital process automation (DPA). With these technologies -- especially with AI and machine learning -- an IPA tool should be able to learn how to adjust and improve the process flow to create an intelligent process. It should be able to learn and improve over time."
So there are choices about how to define it. Too often the definitions lean on the technologies being used (RPA, AI, Machine Learning, Digital Process Automation, or BPM, etc.).
In my own estimation: think of IPA as mission critical processes (or applications to support those processes). Bring all the technical toys to the party - not just the process/automation oriented software, but the modern application architecture tools (event streams, containers, micro services, all the buzz words). Focus on the people the applications support (customers and Internal team), focus on the processes that support the jobs they need done, then design the systems to support those processes.
If you start with the processes and people you'll be able to choose the right tools (technology) to get the work done -
Scott
(ps - obviously, if you find this point of view compelling and need help, our firm - BP3 - focuses on exactly these kinds of opportunities)
Find out what your peers are saying about UiPath, Automation Anywhere, Microsoft and others in Robotic Process Automation (RPA). Updated: January 2026.
Intelligent Document Processing utilizes AI technologies to streamline document-based workflows, reduce manual input, and increase efficiency in data extraction and processing.IDP integrates advanced technologies like machine learning, natural language processing, and computer vision to analyze unstructured data. It automates tasks like data extraction, classification, and validation, helping organizations minimize errors and process documents faster, thereby improving collaboration and...
Let me start by saying that it's just a catchy word, like many others in our industry. Depending on who you ask (there are some good answers here already) you will get a different answer.
I would summarize it as the concept of applying different AI and other data science-related methods to enhance how you typically envision task and/or process optimization initiatives. From how you define and design your future process implementation, to how it will actually perform some of the tasks involved, or how your implementation adapts to changes over time.
So it's a super broad spectrum, where different vendors will pull the concept to their area of expertise.
I'll try to summarize some real-world examples across that spectrum:
- Automated/assisted current process discovery, benchmark and diagnostics, to accelerate your process transformation efforts, and feed that into your process execution stack.
- Process simulation / digital twin simulation, to support strategic decision making based on sophisticated business simulations (like a super "Excel What-if" feature but at company scale) to design optimal TO-BE processes.
- Automated/Assisted task flow design (based on previous topics), to ease implementation of currently manual task sequences into your task/process automation tools.
- Smart/dynamic run-time decision making, versus static, rule-based decision making (like sending a customer request to a different queue based on predictive churn risk, instead of static rules, like customer scoring.
- Process self-healing/adjustment: Use ML to let the process self adjust based on past experience to optimize outcomes. A basic example would be around decision-related thresholds that are adjusted based on past decision outcomes.
- Actually performing tasks automatically, not that easy to automate by "older" approaches. Like validating an ID with computer vision technologies (not just reading the ID, but actually validating it's legit), instead of sending those to a pool of humans for verification.
- Enriching the context data set the process can use to take optimal action, like automatically extracting customer intent from voice/chat, or applying OCR to read a scanned invoice.
- Smart monitoring & management of your process automation platform. An example would be to manage resources dynamically based on learned patterns, and avoid unnecessary resource consumption, or adjust for a process load hike based on past behavior.
I hope this helps understand not the term, but the actual potential behind it, regardless of who's trying to sell you what ;-)
Hi @Evgeny Belenky,
Intelligent Process Automation is basically an upgraded form of robotic process automation. Unlike RPA, intelligent process automation can understand context, learn, and iterate. IPA can also handle both unstructured and structured data and supports some level of informed decision-making. Informed decision-making can further be divided into task level or process level automation. IPA helps organizations to access and analyze unstructured data like images or text that is inaccessible by other means to gain important insights.
IPA can take unstructured data and turn it into structured data for use with RPA technologies. For reasons like these, the technologies are not mutually exclusive but can work together to optimize business processes.
So, while we may be a long way off from robot servants that do our laundry and mow our lawns, robots are already and will continue to play increasingly important roles in our daily lives. Robotic process automation makes work less tedious and allows organizations to scale their operations to provide more value to customers. Intelligent process automation builds upon RPA, giving systems the ability to automate tasks and to learn from them.
For organizations that are new to automation, getting started with RPA technologies is the way to go. They are easy to use and can be implemented with low-code business process management software. Once automation has been introduced into one or more processes, organizations can consider introducing intelligent process automation (IPA) for complex business processes.
Use cases of IPA
Comparing Prices
Companies generally engage in large purchases to offer varied services or products. The cost of these purchases can affect the company's revenue or profits and hence companies tend to research online for making informed decisions. This exploration for right decision making can take a long time and involves complex processes, which is the reason why many companies are now adopting IPA. The right IPA solution can not only compare prices from different vendors but can also compare product attributes and quality. Companies employing IPA can now buy the best services or products at the best prices possible.
Customer Information Storage
IPA can help you store, organize and categorize all types of customer information to ensure easy access to everything. The system can automatically categorize different data sets such as contact information, purchase history, preferences or personal information. The system can display all information to customer support team, sales representatives and similar employees. There is no need to enter this information manually or worry about its accuracy. IPA can always be depended for accuracy in contrast to employees and has a lower error margin. IPA can thus help you reduce repetitive task saving you from unnecessary stress and labor involvement.
Recruitment Process
Recruitment is another prime candidate for implementation of IPA system in large organizations. IPA can aid the recruitment process by helping the HR team to source resumes from a range of online platforms, analyze the exact skillsets, derive value, sort through spams and unwanted applications. It helps the HR team to gain access to the right skills pertaining to candidates with much reduced effort and cost efficiency. IPA not just filters unwanted applications but also helps HR to sort through the right resumes and have access to every application exactly suited to the job requirement. IPA helps the HR team throughout the recruitment process from screening to assessment to final onboarding and administration.
@Shibu Babuchandran Thank you for this explanation. Direct and clear.
I think the market largely has the wrong definition of IPA, but we can turn to HFS for a take on it: "
"Intelligent Process Automation Definition: Intelligent Process Automation (IPA) is the use of technology to allow a business function or part of the operation of a process workflow work automatically. It includes the use of RPA, BPM suites, Remote Desktop Automation, screen scraping and custom scripting and related technologies." (HFS, 2018?)
I think it is fair to conclude that
Tech Target: "IPA is designed to assist human workers by doing manual, repetitive and routine tasks that were previously performed by humans. Technologies combined in IPA include robotic process automation (RPA), artificial intelligence (AI), machine learning and digital process automation (DPA). With these technologies -- especially with AI and machine learning -- an IPA tool should be able to learn how to adjust and improve the process flow to create an intelligent process. It should be able to learn and improve over time."
So there are choices about how to define it. Too often the definitions lean on the technologies being used (RPA, AI, Machine Learning, Digital Process Automation, or BPM, etc.).
In my own estimation: think of IPA as mission critical processes (or applications to support those processes). Bring all the technical toys to the party - not just the process/automation oriented software, but the modern application architecture tools (event streams, containers, micro services, all the buzz words). Focus on the people the applications support (customers and Internal team), focus on the processes that support the jobs they need done, then design the systems to support those processes.
If you start with the processes and people you'll be able to choose the right tools (technology) to get the work done -
Scott
(ps - obviously, if you find this point of view compelling and need help, our firm - BP3 - focuses on exactly these kinds of opportunities)
In simple words, anything which needs OCR capabilities or to deal with processing of unstructured or semi structured data is IPA, rest it is all RPA.
Rest, it is all about jargon - hyper automation, 360 automation etc....