Elevating IT with R&D: ChampionX CIO Alina Parast on AI, Cybersecurity, & Talent | Technovation 795

Thank you, Peter, and great to be part of your conversation.

Yeah, absolutely. So ChampionX is a global company and we operate in the energy sector, and we have several work streams of focus within ChampionX. So we provide chemistry solutions, we provide artificial lift solutions, and we also provide a number of highly engineered solutions and equipment in the energy sector to help companies to provide oil and gas to countries' populations in a safe, efficient, and sustainable manner. And recently, we also expanded in the emission monitoring, and we also offer digital solutions to our customers as part of our product as well as our services. And we operate in over 50 countries around the world. We have about 8,000 employees, and we're a global public company.

So thank you, Peter. So I love my role first of all, and I enjoy every minute of being an IT leader for ChampionX. It's a diverse role. My primary focus is to make companies successful and to serve our customers and participate in our communities and improve lives of our employees and stakeholders. So what I'm responsible for today is for global infrastructure and operations, which is more or less traditional role. I'm also responsible for global security, for cybersecurity for ChampionX. I'm responsible for enterprise platforms and applications, as well as data and analytics, and primarily data provisioning and getting into some newer areas such as AI. And I partner with all our business areas such as supply chain, finance, regulatory, as well as our business units that focus on our customers and deliver products and services.

Yeah, so while we've always supported our R&D employees in a traditional sense by providing tools and platforms and everything, but this is kind of expanding and it's really exciting to me because it's a new area to really get deeper involved and make a difference in some new areas where we haven't participated in the past. So we began partnering with the R&D team at the cusp of AI exploding, so it kind of offered a new avenue for us to partner with research and development. And we're looking at specific areas to revolutionize R&D, but maybe more pragmatic is to really introduce continuous improvement and acceleration and automation to a field of R&D. So R&D is science, and in science you do research, you do development, you do experimentation. There's a whole process involved. So where we are getting involved is looking at the process of R&D and where we can identify use cases where AI can be applied to accelerate that and automate that, and to get away from maybe a long duration of experimentation and simulation and discovery. And it's absolutely exciting to step into the world which I worked in R&D in the past as part of my career, but not actually from the IT enablement perspective. And obviously accelerating discovery of new chemistries, new products using AI models is something completely new and very motivational to me and my team as well.

So before we even get into AI, what we did in partnering with my peers in R&D, first thing you learn is that understanding and knowledge of artificial intelligence varies across the board. You have scientists, you have engineers, some know more, some know less. So the first step before you even step into any use cases is education. It's so vital and critical. So where we begin with R&D is to really educate and bring everybody to a level of truly understanding what AI can offer. So it starts at the rudimentary level, just ABCs of AI, and then diving deeper into saying what potential, how could we implement it, what potential use cases you can think about. One difference I think from prior technologies and what we're learning through that is that in using generative AI or any AI in R&D or any other field, the subject matter expert is really at the center. And we are educating members of our R&D team to understand that in order to apply AI, you really need to also understand that you are the expert. And one feature of generative AI is to produce new output, in this case to produce potential new chemistries, new molecules. Only you can really evaluate: is AI hallucinating or is it really a valuable output? So that's one of the areas that we're exploring with them. Can we take the process that typically is very prolonged and use healthcare and pharma as an example? They're much further along. And can you use the data to accelerate creation of new chemistries outside of the lab, for example, through simulation and through application of AI techniques? The other area, which is more traditional AI, is partner with R&D, bring their data into our enterprise data lake, and use machine learning to do prediction and forecasting, also applied from R&D to supply chain in terms of what chemistries can be applied to certain customer problems and how do these chemistries map into our products. So there are many different applications that I think need to follow learning and educating our scientists and engineers so then they can step back and see opportunities, then bring the data, find a home for data to make sure—because AI without data is really not meaningful—so bring the data, secure the data, and then apply AI tools to the data to potential business problems or use cases that they bring to the table, and then they have the experts validate it. But that's kind of where we are with R&D, and I think it has so many opportunities. And also educate them to look outside of their space, look at pharma, look at healthcare, and it's directly related because the use case has a very similar discovery of new drugs, discovery of new chemistry, automation of the process of the research literature research. Automation and AI is amazing at literature research and accelerating that piece of it.

Yeah, I gave them one example. If I, as an IT professional, a technology professional, am not a chemist, only they can tell. So I gave them an example using generative AI to create a molecule for cherry juice and snowflake, and it spit it out for me, but I can't evaluate the validity of that. Only they can. And I think that kind of rudimentary to complex education is very important.

So first thing, also may sound very rudimentary, but something that we have, it's the very first step that we did, is educate people particularly around our IP. You can't just start populating large language models that live in public space. You may think that it's so basic, but it's part of that overall education. The minute you put your formulas, your IP into public large language models, it belongs to everyone. So that was the first step: to work with businesses, with R&D, and really get that point across. The next step would be, so then how would you leverage AI and be safe and secure at the same time? My earlier point is AI is based on data, so we need to find a safe and secure home for our data before we apply AI. And so when we looked around, we are very—we leverage Microsoft technologies quite a bit. There are many others, we just happen to use Microsoft. And first thing we did is when the whole explosion of AI occurred, we already do quite a bit. We have cloud, we have enterprise data lake, we use IoT technologies within the Microsoft space. And then speaking with our partners at Microsoft, we also saw that OpenAI has been incorporated and integrated into Microsoft Azure Cloud for some time, and they're continuing to invest in that. So we said, well, now first step is we can use data that's within our four walls, within our own Azure cloud, and then apply AI technologies to that data. So as I mentioned earlier, first is what is the home for your data? The home is our enterprise data lake. So we can bring more—whether it's research and development data or supply chain data or financial data—we can bring it into our data lake, or we have a lot of it already, and then leverage AI capabilities that are natural to large language models and applied. So our first use case actually was: can we apply natural language against business data that lives in the enterprise data lake? So we're all familiar with creating reports and analytics and all, but what we wanted to do is simply ask questions against the data that lives in our data lake. And we partnered and had a number of sessions doing proof of concept, and we demonstrated that you can apply natural language and ask your business questions against your data that lives in your data lake. And it was a kind of a turning point, I think, for some of our engineers, our business analysts that are used to more of a structured programming, and here they didn't have to do much of that. I mean, once they had data and we did some modeling and applied what our AI already has built-in capabilities, we could ask, 'What would be my margin if I do this?' instead of writing typical reports and analytics. So it was eye-opening, but people had to adjust their thinking on how to go about that. And then of course comes the security of securing the data. Just because it lives in a data lake, maybe it's secure inside, but now you want to preserve some of the integrity of the pricing, of your costing, and everything. Not even everybody inside should be able to access it. So we have to incorporate that into when we apply AI.

So cybersecurity on the surface seems like an unlikely area for having fun, you know. But we all experience corporate training and other training online, and after a while it becomes dry, if not not exciting. And we looked at our—just like any company, we trained employees annually on cybersecurity, we do simulations and stuff, but results, while improving, I would say not fully engaging. So we looked at other avenues on how we can improve. And we don't subscribe to the philosophy of being characteristic on the cybersecurity. I mean, we want people to really internalize that cybersecurity is something that doesn't just belong to a small team in the corner that does cybersecurity and applies it, but it's everybody's responsibility. But it also then needs to be a little more fun than your typical training. So we evaluated various options, and we came across a partner of ours that produced mini-series. And I love mysteries, and I love especially English mysteries. And this was basically done in a very high quality style, it resembles your typical British mysteries and murder mysteries, but it's done all around cybersecurity and focused on getting employees engaged and watching. So we introduced season one that had a bunch of episodes, and we had great feedback. We're on season five, and you can re-watch and rerun, but it's a better way of learning, I suppose. And we brought some speakers from the same partner who approached cybersecurity in a little different fashion. And they would use an example and say, 'Just because you want to be healthy, you don't necessarily exercise, you still eat sugar.' So we want to get the same idea of getting you to the point where it's your internal goal to be safe, healthy, and secure. So I think that's the philosophy. And of course engaging our executive team, engaging our managers—they do an amazing job also of improving employee focus on cybersecurity. And we're doing something later this year is to have a simulation for our CEO and cybersecurity, and our executive team engaged in cybersecurity gaming approach.

So early on, as we became a public company and we focused on cybersecurity, we were building technology grown up to enable a new company—even though we're not new by age, quickly became a new public company. We had to bring many new partners and vendors into the picture, and it's always important, particularly on the technology side, to ensure that they're capable, they have good practices, and their cybersecurity is safe. So you have traditional methods: you ask for their audit reports and all that good stuff. But early on, we started exploring new ways. A year and a half into that, just like you can think of a person having a credit score, a person could have a cybersecurity score. And I think it's becoming more mainstream, but going back a year and a half, we said, 'Can we—we spend a lot of time just manually asking for their architectures, their policies, their processes.' And I think we have found some good partners who significantly simplified and standardized and moved us closer to that cybersecurity score. So today, in addition to traditional methods, we use partners that run extensive assessment from the outside and produce a cybersecurity scorecard for potential vendors. And then we supplement with other things, but it definitely improves standardization, accelerates the process of assessing cybersecurity posture of our partners. We then turned around and said, 'Well, if they can do that for our partners, we can do it to ourselves.' So we started running a cybersecurity scorecard against ourselves, which definitely uncovered some areas we needed to improve. And our latest step is really focusing on our investors. So we partnered with yet another company that focuses on investors, and we work with them on improving our investor-focused cybersecurity score. So just a different way of approaching cybersecurity, I guess.

I think as we look into the future, in particular for IT professionals, technology professionals, what's exciting for me—and I think an opportunity for others—you step out of the traditional. You still have to do traditional stuff, don't get me wrong. You have to do your ERP, you have to do your networks, you have to do that. But it's exciting as the technologies change around you, you step into learning about the business you're in or industry you're in. And I always say you triangulate with somebody like we used example of pharma, and you learn about your business, and you look at the technologies in networking and cloud and artificial intelligence. And even traditional things like ERPs are moving way to the next generation, and you're looking at opportunities to—how do they triangulate, how they map to the way you run your business. So even ERPs that are now moving into SaaS, the last ones I think of, I mean they offer opportunities to leverage them for your decision making, for business decision making, for supply chain visibility in a way that you couldn't do it before. And I think it's that with all the platforms that the businesses are using, kind of moving in the space to be an equal level in a software as a service with data available on multiple aspects in the data lake, you can do more, faster, with less.

Well, in every situation, obviously the teams and your business partners are amazing, so you have to continuously learn from them and learn from your partners and your vendors. I mean, there are opportunities to get these nuggets, these amazing ideas from the ecosystem of people you work with. And never stop learning. But the other principle that I always apply and I share with my team is less is always more. We don't need 55 tools or 100 platforms. So simplicity—and I kind of always look at Apple, you know how that simplicity was always the strategy. I mean, the simplicity within being an effective partner to the business in IT is: if you can't explain it, you can't support it, you can't roll it out. So keep it as simple but sophisticated as you can. And I think that we strive for that in everything in our networking and our cyber and applications. Less is more, for me anyways.

Thank you, Peter, for having me. I appreciate it.