AI Insights From the CIO of Automation

A lot was learned about GenAI in April 2023. This was a huge event where much emphasis was on automation. Both the CIO and AI angle was touched upon which can thereby be implemented for 2024. Eminent speakers shared their thoughts on the same. Amazingly, GenAI may change the game for all of them by rewriting the rules of what’s possible, in addition to helping.

I have listened to many Tech speakers but this one is special as it gave me a new insight altogether.

The hosts, Gabriel Carrejo and Micah Smith are joined in this episode by Sumit Johar, CIO of Automation Anywhere, to talk about the steps needed to implement AI.

For chief information officers and IT departments, Sumit will impart wisdom and offer useful advice. Think about the Chief Information Officer (CIO), Chief Technology Officer (CTO), Chief Data Officer (CDO), Chief Security Officer (CISO), Chief Enterprise Architect (CEO), etc., and how their interactions have changed since the advent of Gen AI. Whose decision is it to own?

Embracing generative AI (GenAI) is not a choice but a strategic need for every modern organization, as Sumit shared honestly as a chief information officer who has seen the revolutionary impact of GenAI personally. The goal of IT executives is to maximize efficiency in all three areas: price, size, and expertise.

Being at the forefront of the automation business, working with GenAI has been like riding a massive wave of innovation. Adaptability, bravery, and initiative have been essential for making it through this deluge of change. More crucially, though, it has made possible a plethora of possibilities that were beyond our reach.

Read: The Top AiThority Articles Of 2023

They are at the forefront of IT, therefore they can’t afford to sit on their hands and wait for change; they have to make it happen.

First Chapter: Where Automation Anywhere’s Generative AI Journey Began

Since they were pioneers in the field of Intelligent Automation, they were caught off guard by the introduction of GenAI.

They were among the first businesses to use GenAI when they embarked on this path. Many people were worried and confused at first. Looking back on the early days of GenAI adoption (which seems like it happened just last year), enterprises were naturally wary of the potential benefits and dangers of the technology. Concerns and doubts around the potential benefits that GenAI could bring to companies. It mirrored the development and spread of cloud computing in many respects.

However, they were not going to sit on their hands and watch.

Their culture and the nature of our business made it mandatory. Here at Automation Anywhere, they believe in constantly pushing the boundaries and trying new things. “Customer Zero” is our operational program to test on customers, get insight, and enable the creation of tried-and-true solutions to provide our customers. It’s a formalization of our commitment to being on the front lines. So, it’s no surprise that they were excited to launch GenAI as soon as it was available. The sheer magnitude of the situation was beyond their comprehension, though.

Chapter 2: Facing disruption and embracing risk

Privacy and security were the most frequently mentioned worries, especially at the beginning. The safety of our data is your number one priority as chief information officer when implementing new technologies like GenAI. Preventing the loss, misuse, or unauthorized disclosure of sensitive information is of the utmost importance.

The word disruption is used frequently and has a large impact. Here, though, I’m at a loss for words to describe the revolutionary change that GenAI has wrought in every industry. It was a game-changer for those of us working on intelligent automation from the get-go. Inevitably, panic sets in as a first response to such disturbance. But they weren’t fooled into thinking they had a crystal ball that could see where this technology or the market was going. True bravery is not being fearless but rather being able to move forward despite fear; they truly had no choice but to model this quality after us.

They were aware that GenAI posed a risk to the security of company data. What was our response, then? Until we had a better grasp of the consequences and could come up with tangible remedies, we opted to deal only with public or non-sensitive data regarding data privacy. There were new mysteries to discover at every curve along the way. Not to mention that the majority of them still lacked satisfactory responses. There was a choice: sit tight and hope for answers to materialize, or get to work discovering them yourself. Hint: In hindsight, I can assure you that there was a solution to every issue and roadblock we faced.

Chapter 3: Resetting our plans and expectations to match the opportunity

They removed their gloves and dug in, using both hands, to reach the bottom. To determine where GenAI could make a difference, they looked at the fundamentals of automation, the meat and potatoes of our business and platform. It should come as no surprise that we discovered the solution all over.

A few months later, after they had conducted an additional study on our behalf, we came to fully appreciate the magnitude of the potential and the possibility presented by GenAI. So, we have arrived at a stage of realization and reset in our GenAI journey. First and foremost, when faced with an unforeseen event, it is essential to acknowledge your current location and the gravity of the situation, according to Navy SEALs. All of our plans had to be shelved when they discovered we needed to start over. The claim that GenAI revolutionizes the industry is borne out by their own experiences.

As an example, unstructured data used to necessitate pre-training before it could be utilized to automate business processes, and inputs to automation were limited to pre-structured information. Now, there’s no requirement for pre-training; inputs could be data and information in practically any format. Furthermore, GenAI allows for the possibility of outputs in any format. A picture or a presentation slide set might do. Contextually produced rules for the process, including the automation itself, could one day reference past data through natural language communication with GenAI, eliminating the need for definition and programming.

The current limits of automation were blown wide open by GenAI. They started to reevaluate our entire company and product strategy. Capitalizing on the enormous potential and acceleration, existing plans and investments were adjusted. As word of this change got into their company and product development processes, new plans meant new OKRs. The same was true for their technology stack. Some technology and tools would be necessary, while others would be out of our price range. For GenAI, which tools are they looking for? How might GenAI be integrated into routine duties to facilitate this new way of working?

Chapter 4: Embracing the art of the possible

They managed to hold off on making assumptions and smoothing things over until they could see a foundation taking shape, which led to the ongoing temptation to begin laying that foundation. However, that was the wrong way to go. A base cannot be constructed out of nothing. Discovering more, and learning about what we didn’t know, was crucial to completing the picture. Thus, we embarked on a period of serious experimentation.

They were so engrossed that they launched a company-wide tournament called “Demo Royale” to channel the enthusiasm of friendly rivalry and discover immediate, practical uses for GenAI that our employees might identify by drawing on their knowledge in certain areas. Nearly half of their staff took part in the outing. With over two hundred comprehensive use cases submitted, selecting winners was an arduous task.

At the end of the day, we had to break the tie and provide $10,000 to the top finishers. Through everyone’s exploration and experimentation with this technology, we emerged with a plethora of ideas, boundless energy, and a clear picture of what was possible. It was only the beginning of what was to come.

Chapter 5: Building our foundation with the right GenAI model

Sure, how can they have put this plan into action? Finally, in the trek through GenAI, we had reached this far-from-trivial question. Now that we have found the fundamentals for making GenAI operational, they laid the groundwork.

It was necessary to assemble a team of GenAI specialists to bring our concepts and application cases to life. What is the best place to hold this skill set? In your company, which department or team would serve as GenAI’s nerve center? It was easiest to begin with our center of excellence, our automation team. The CoE is constantly working on this project, learning and improving their GenAI skills by doing. They are in the trenches every day.

They had also reached one of the journey’s major questions: which LLM (large language model) should we fund? For this, they contacted an old business associate and ultimately settled on two models to implement. The importance of training the model on the correct data set to enable it to augment our enterprise knowledge was another lesson we picked up.

To state that training and tweaking your selected LLM is the most complicated aspect of this story would be an understatement.

Applying GenAI outside of its training domain of expertise increases the likelihood of accuracy issues and “hallucination” dangers. It is critical to train your selected model(s) using your enterprise data, as our experience has demonstrated. Verify your model, indeed. Retest if necessary.