Achieving automation at scale: why CEOs should care

And it’s happening to CEOs. They just don’t know it yet.

Executives are not always calm, cool, and collected. It just seems that way. And probably is that way most of the time. But here is an excerpt from the minutes of the ongoing committee meeting in the CEOs head:

Wait- shouldn’t the word be scalability? I thought that was a routine value added service, like free fries if I bought the jumbo burger. Scalability is all upside, right? And I am not required to have this ability to scale- that comes with it, doesn’t it? Now we are achieving automation at scale. The scaling part sounds less automatic all of a sudden. Where is my IT guy? Why do I have to care? What do I have to do? What do I need to know?

There is a subtle change in the wording, which represents a not so subtle innovative advance in robotic process automation, (RPA). Application of advanced AI with machine learning, deep neural networks, and data analytics is useful in RPA, just as it has been in NLP. In the process automation arena, what is now called intelligent process automation (IPA) is a far cry from first generation RPA.

"RPA is about building bots that automate routine, repetitive tasks. IPA goes beyond that, leveraging AI to automate more complex tasks that require decision-making, analysis, and collaboration between humans and bots. With IPA, we can achieve not only more efficient operations, but also new insights and opportunities for innovation that were previously inaccessible." - Mary Lacity, Professor of Information Systems at the University of Missouri-St. Louis and co-author of the book "Robotic Process Automation and Risk Mitigation: The Definitive Guide" (2020).

The key differences between traditional RPA and IPA lie in the complexity and scope of tasks that they can automate. While RPA is effective at automating repetitive, rule-based tasks, IPA can automate more complex tasks that require decision-making, analysis, and collaboration between humans and multiple bots. However, IPA still faces challenges in terms of identifying automation opportunities and the high cost associated with developing and maintaining these more advanced systems.

The 2020 Deloitte study conducted on RPA revealed so many great statistics. The following percentages were the percentage of respondents that claimed such specific benefits from their own RPA projects, each of which varied widely.  

Bain & Company did their own study on RPA, looking at it from a different angle. The biggest surprise? 66% of respondents at the time had never even heard of RPA, and had zero understanding of its potential value to their firm. This is not good.

Really, it’s about the opportunities for automation. That’s the key. If you figure out how to find, characterize, and catalogue these opportunities properly, you’ve captured some legitimate % of the value already.

To address the challenges of identifying automation opportunities, research efforts have focused on using structured and unstructured data from process logs to identify candidate tasks for automation. However, more work is needed to scale these efforts and provide recommendations for suitable IPA templates or AI models.

The cost of building and maintaining IPAs is also higher than that of traditional RPAs, as it requires data preparation, feature engineering, and ongoing retraining of AI models. To mitigate these costs, efforts can be made to decrease the effort required to develop IPAs, enable them to be reused for different types of processes, or use them to replace or augment different customer tasks.

Intelligent Process Automation (IPA) and what it requires

You may feel like with more primitive RPA, you have spent your time with training wheels. You may be confident that you are ready to balance the bicycle yourself without these mechanical failsafe devices. The problem is you're riding inherently on a different road. The bike with the training wheels for more primitive RPA is only used on quiet country roads. Putting on the gloves for full blown IPA is only needed on the busy Interstate exchange of a large city. You're going to be the only one on a bicycle. Don't be too hasty because the stakes are high…

Required capabilities to develop, install, and implement IPA include:

1. Data preparation
   The ability to identify relevant data, clean and transform it into a format that can be used for IPA.
2. Feature engineering
   The ability to extract relevant features from the prepared data that can be used to train AI models.
3. AI model development
   The ability to build and validate AI models that can automate complex tasks and decision-making processes.
4. Integration capabilities
   The ability to integrate IPA solutions with existing systems, applications, and processes to ensure smooth and efficient operations.
5. Process monitoring and optimization
   The ability to monitor IPA solutions and continuously optimize them to improve performance and efficiency.
6. Change management
   The ability to manage changes to business processes, data, and systems that may impact IPA solutions.
7. Risk management
   The ability to identify and mitigate potential risks associated with IPA implementation, including cybersecurity risks, data privacy risks, and regulatory compliance risks.
8. Training and support
   The ability to provide adequate training and support to end-users and stakeholders to ensure effective use and adoption of IPA solutions.
9. Scalability
   The ability to deploy and maintain IPA solutions at scale, with minimal disruption to existing operations.
10. Business process expertise
    The ability to understand and analyze complex business processes and identify opportunities for IPA implementation that can drive business value and innovation.

Why all the worry about scaling IPA projects appropriately?

"Scaling is more critical with IPA than RPA because IPA involves a more holistic approach to automation, encompassing the entire lifecycle of a process and often requiring the coordination and collaboration of multiple solutions. While RPA is focused on building individual bots to automate repetitive tasks, IPA aims to automate more complex tasks that involve decision-making and analysis, which requires a broader set of capabilities and a more integrated approach. Improper scaling of IPA solutions can result in inconsistent outcomes, poor performance, and limited scalability, which can have significant consequences for organizations." - Dr. Satish Joshi, Director of Data Science and Engineering at Western Digital.

Improper scaling with IPA projects can result in a number of risks, including:

1. Decreased efficiency: If IPA solutions are not scaled properly, they may become inefficient and slow, leading to increased processing times and reduced productivity.
2. Higher costs: Improper scaling can result in higher costs for IPA development and implementation, leading to lower return on investment (ROI) and financial losses.
3. Inconsistent outcomes: Inconsistent scaling of IPA solutions can lead to inconsistent outcomes across different processes or business units, resulting in lower quality outcomes and reduced customer satisfaction.
4. Poor performance: If IPA solutions are not scaled properly, they may not perform as expected, resulting in errors, delays, and other performance issues.
5. Security vulnerabilities: Scaling IPA solutions improperly can lead to security vulnerabilities, exposing sensitive or confidential data to unauthorized access or attacks.
6. Compliance risks: Improper scaling can result in compliance risks, as IPA solutions may not be able to comply with regulatory requirements or company policies.
7. Resistance to change: Improper scaling can result in resistance to change, as employees may be hesitant to adopt new IPA solutions that are not properly scaled or integrated with existing processes and systems.
8. Maintenance issues: If IPA solutions are not scaled properly, they may require more maintenance and updates, leading to higher costs and more downtime.
9. Limited scalability: Improper scaling can limit the scalability of IPA solutions, preventing them from being expanded or adapted to new business needs or requirements.
10. Reputation damage: Improper scaling can damage an organization's reputation if IPA solutions fail to deliver expected outcomes, leading to dissatisfied customers or negative publicity.

What are the benefits of proper scaling in IPA?

The good news about scaling with IPA:

1. Efficiency
   Scaling IPA solutions can lead to significant efficiency gains by automating complex tasks that were previously performed manually. This can lead to increased productivity, reduced operational costs, and faster turnaround times.
2. Increased ROI
   By scaling IPA solutions, organizations can realize a higher return on investment (ROI) by spreading the costs of IPA development and implementation across a larger number of processes or business units.
3. Consistency
   Scaling IPA solutions can ensure consistent and standardized execution of tasks across different business units, processes, and geographies, leading to better quality outcomes and reduced errors.
4. Innovation
   By scaling IPA solutions, organizations can create a culture of innovation and experimentation, driving continuous improvement and new opportunities for growth.
5. Flexibility
   Scaling IPA solutions can provide organizations with the flexibility to adapt to changing business needs and requirements, as IPA solutions can be quickly reconfigured or redeployed to different processes or business units.
6. Competitive advantage
   By scaling IPA solutions, organizations can gain a competitive advantage by being able to deliver higher quality products or services, faster and at a lower cost than their competitors.
7. Customer satisfaction
   By scaling IPA solutions, organizations can improve customer satisfaction by delivering faster and more efficient services, and by providing real-time access to information and support.
8. Risk management
   Scaling IPA solutions can help organizations manage risks by automating tasks that are high-risk or prone to errors, and by providing real-time monitoring and alerts to identify and address issues quickly.
9. Regulatory compliance
   By scaling IPA solutions, organizations can ensure compliance with regulatory requirements by automating tasks that involve sensitive or confidential data, and by providing a complete audit trail of all IPA activities.
10. Future-proofing
    Scaling IPA solutions can future-proof organizations by creating a foundation for digital transformation, enabling them to leverage new technologies and capabilities as they emerge.

Takeaways

Here is a list of key thoughts:

1. IPA (Intelligent Process Automation) is a type of automation that involves the integration of AI (Artificial Intelligence) and RPA (Robotic Process Automation) to automate more complex tasks that require decision-making, insights and analysis, or the coordination and collaboration of multiple solutions.
2. IPA aims to streamline and optimize business processes, reduce costs, and improve efficiency by automating tasks that were previously performed manually.
3. To develop, install, and implement IPA solutions, organizations need to have a range of capabilities, including data preparation, feature engineering, AI model building and validation, and ongoing maintenance and updates.
4. Scaling IPA solutions requires a comprehensive approach that involves the integration of multiple solutions, ongoing monitoring and optimization, and the ability to adapt to changes in business processes and data.
5. Improper scaling of IPA solutions can result in decreased efficiency, higher costs, inconsistent outcomes, poor performance, security vulnerabilities, compliance risks, limited scalability, and reputation damage.
6. To overcome the challenges associated with IPA, organizations need to have a clear understanding of their business processes, identify automation opportunities, develop a strategic roadmap for IPA adoption, and continuously monitor and optimize their IPA solutions.
7. While IPA presents significant opportunities for organizations to streamline and optimize their operations, it also requires a significant investment of time, resources, and expertise to develop and scale effective solutions.

Automation - it’s here. Robotic process automation, (RPA), is now a thing. Accept it. Companies are using it, power users are right-clicking stuff they probably shouldn’t be, and developers are receiving requirements documents that say stuff like, “create a macro that generates a pivot table of sheet1 and sheet2.” This is happening.