Status Go: Ep. 249 – The Rise of Intelligent Automation: From RPA to Gen AI | Peter Xu

Welcome to Status Go, a podcast about what it takes to thrive among the challenges of enterprise technology. My name is Paul Wooten and I am the sales enablement lead here at InterVision.

Today I have the pleasure of talking with Peter Xu. Peter is our lead Gen AI solutions architect here at InterVision. But it’s not just InterVision. Peter has over 25 years in the IT and software industry, including spending over 20 years at IBM, two years at AWS, before moving to consulting companies and joining us here at InterVision. Peter brings a deep technical knowledge and go-to-market industry expertise in automation software stacks, both new and old.

He has knowledge from RPA and BPM to what we’re going to focus on today, talking about intelligent automation with Gen AI. Peter, welcome to our podcast.

Thanks for having me, excited to be here.

Peter, let me start by asking you a couple of questions. Now, in the past, your industry knowledge, you have worked extensively with automation technologies like BPM and RPA. Now, talking about RPA, what was the original promise of RPA, and where did it fall short?

Absolutely, Paul. So from my past experience, I think the original promise of RPA became very attractive. It was envisioned as a tool that would replace human effort on repetitive tasks, thereby boosting efficiency, reducing costs, and seamless integrating with the existing system. In practice, RPA proved to be highly effective for small, well-defined tasks. However, its limitations soon became very evident.

RPA systems lack the adaptability to learn and evolve alongside changing business needs, which meant they struggled to scale effectively across departmental solutions. Moreover, the cost and maintenance of this system sometimes escalated, making them more expensive than initially anticipated. So this experience highlighted that while RPA laid a solid foundation for automation, it requires further enhancements to fully meet the dynamic nature of modern enterprises. Now, of course expense was one of the big limitations of RPA, but when you think about the technological limitations of RPA, when it came to workflow automation and improving productivity, what were those biggest technological limitations that you saw? Yeah, think RPA’s biggest technology limitations stem from its rule-based design.

While it was effective for structured tasks, it struggled with unstructured data and making complex decision making, limiting its ability to adapt to change. Additionally, its reliance on screen scraping or UI-based automation meant that every minor change in user interface can break the workflow, leading to increased maintenance burden. Thirdly, although RPA is often marketed as a low-code solution, it still demands significant customization, ongoing IT support, which ultimately undermines its potential to boost productivity.

It has a lot of challenges right there. Now, just think about those challenges that RPA faced. What makes Gen AI fundamentally different? So why is it succeeding where those past approaches have struggled or even failed?

Yeah, you’re right, Paul. Generative AI definitely represents a fundamental shift from traditional rule-based RPA by offering a truly adaptive approach. Rule based relies on predefined steps, Gen AI leverages advanced large language models to work with unstructured data such as documents and images and supports complex and dynamic decision making. This allows it to automate not just routine tasks, but also knowledge work and dynamic workflows. As a result, Gen AI is redefining employee productivity and the business process, rapidly accelerating adoption across industries and reshaping roles within the enterprise as well.

I understand that. One more thing. Today, a lot of companies are feeling pressure to do more with less. So how is Gen AI helping them to achieve this?

Yeah, absolutely. In today’s competitive environment, everybody asking to do more is less. is proving to be the real game changer. Unlike traditional BPM, RPA, or even older AI technology, Gen AI delivers the best outcomes cheaper, faster, and better, reducing the need for heavy custom scripting and high maintenance setup. It integrates seamlessly with APIs, databases and the existing business systems, reducing overhead, and streamlining operations. By automating routine tasks across departments, Gen AI not only boosts employee productivity in HR IT, it can also enhance content report generation in sales, marketing, and finance, while helping operation leaders simplify complex processes across enterprises.

Yeah, I hear a lot of examples from my sales team about how they’re using AI to automate a lot of their work. And those really resonate. So what are some other real world examples of how Gen AI can be used today to automate work that you’re familiar with?

Yeah, happy to share some of them, Paul. Gen AI is already transforming work across various roles. For instance, you probably heard about a tool called Perplexity, which has streamlined customer and industry research, by quickly analyzing vast amounts of data. Another tool I like a lot is Notebook LLM, which could automate production of marketing enable material or even generating podcasts like this. In the software development world, GitHub Copilot assists engineers by offering intelligent code suggestions and automating routine tasks, significantly speeding up the development process. Additionally, AI-powered video analytics also revolutionize how businesses analyze video content, enabling faster and more accurate extraction of insights from multimedia data.

Now, even though there’s all these options out there, there’s these real-world examples of success. Now, a lot of companies still say they’re not ready for automation with AI. So what would you say to those companies that think they’re not ready for automation with AI? It’s too hard to implement, their data isn’t ready. So what can they do to get started?

Paul, I think that’s a good one. I hear it a lot from customers. For those concerned companies, I think the first step is to recognize that perfect data is not always a prerequisite. In fact, modern AI model is adept at handling unstructured and incomplete data, allowing you to start reaping the benefits even with a less than ideal situation or data set. I always suggest customers and companies begin by innovating the art of possible, for your enterprise. Consider the transformative impact AI could have on your operation. Then start small by implementing intelligent search tools or even automate some simple tasks. As you build confidence and see the results, gradually expand to full end-to-end workflow automations.

Most importantly, also, invest in training and enabling your employees to ensure that they’re well equipped to leverage those new technologies.

Very key. Enable them, train them, keep them going, because AI is here, and I’m pretty sure it’s going to be here to stay for a while. Let’s take a look to the future right here. How do you see the role of intelligent automation applications evolving in the next couple of years?

Yeah, I’m very optimistic, and firmly believe over the next couple of years, intelligent automation is set to become even more integral to business operations. With LLM and token prices dropping fast, also the rise of open source models, AI technology is becoming more accessible and affordable. This evolution means that AI will no longer be a niche add-on, but will form the very foundation of business applications.

As a result, the competitive advantage will shift from merely adopting AI to the excellence in the application layer. Focus on how AI is integrated to deliver unique value-driven vertical solutions.

I love your optimism there. I think it’s very, very important to keep that going and just be open-minded. I use AI a lot. Everybody uses it. The more you use it, the more you’re going to keep using it at this point. Now, one last question right here. Our listeners are probably familiar with Gen AI RAG. It’s retrieval augmented generation. as we talked about, AI is constantly evolving. What is that next wave in Gen AI that you see and how do you see it being different?

You’re absolutely right, Paul. AI is a fast moving space. The next wave of Gen AI is marked by the rise of the AI agent. A significant evolution beyond RAG.

While current RAG enhances AI by enabling real-time knowledge retrieval, making response more contextual, rich, and accurate, AI agents take the concept further by actively taking action. These agents are designed not only to retrieve information, but also to plan and execute multi-step workflows, automate decision-making processes, and trigger actions across various enterprise applications.

In essence, AI agents transition from being merely an assistive tool to become an autonomous system that can manage entire workflow end-to-end with minimal human intervention. This shift means AI will increasingly underpin business operations, handle complex and dynamic tasks that require planning, adaptation, and real-time response.

This evolution is poised to drive remarkable gains in efficiency, innovation, and operational agility across industries. I’m personally thrilled about the exciting future it promises. Peter, I, for one, welcome all these changes. This is great. I’m excited to see what the next few years are bringing us. And, you know, bring myself up to even more speed now that I have more background on what Gen AI can do.

Peter, I want to thank you very much for taking time today. Very informative. I’m sure everyone has great time listening to us today. Again, I want to thank everybody for listening. Remember to subscribe wherever you get your podcasts and visit intervision.com to learn more. Thank you all very much. We’ll see you next time. Thanks, Paul. Thanks, everyone. Thank you, Peter.