Today we welcome Todd James, Chief Data & Technology Officer at 84.51º, to The CXO Journey to the AI Future podcast. He joined 84.51º in August 2021, where he paved the way for continued growth, building on its rich data, science, and technology capabilities. Previously, Todd spent 15 years at Fidelity Investments where he held a variety of key strategic leadership roles. As an innovative leader, Todd built the global data and analytics organization for Fidelity’s Workplace Investing and Health Care business units. He also led efforts to modernize servicing and operations through applied artificial intelligence and automation and directed Fidelity’s Cross-Enterprise AI Center of Excellence (COE).
Question 1: First Job: Could you talk a little bit about your background and how you got to the position where you are now?
We’re a retail data science organization that is part of The Kroger Company. As I like to think about it, what we do is we help Kroger run a media operation that goes to market.
We also have an insights business that helps CPGs and partners understand the path to purchase, and a venture capital fund called PearlRock Partners that invests in emerging food services businesses. Additionally, we’re the advanced analytics arm of the parent company of Kroger. So when it comes to AI and advanced analytics across Kroger, we’re tasked with driving a lot of those results across the organization.
It’s interesting that throughout my career I’ve moved between business leadership roles and technology leadership roles. So while I was at Fidelity, I held a variety of responsibilities from running business units to strategizing the organization. The tail end of my tenure there was really around building out data and advanced analytics capabilities for the workplace, investing in healthcare businesses, and setting up the global capabilities for that. So I had a great opportunity working at Fidelity to do a lot of different things, including help build out the advanced analytics and AI journey that they’re on today.
Question 2: Generative AI: Everybody’s talking about it. You’ve spent a lot of time with data and analytics for years. Could you put a little context around Generative AI? How is the Gen AI movement different?
It’s different, but a lot of the capabilities aren’t necessarily new. There are things happening today that we’ve already been doing for 5-7 years. The difference is that what was very difficult to do (and on a limited scale from a usage POV), has become a lot easier to implement. It’s far more democratized.
The capabilities are greatly expanded relative to some of the natural language solutions that we were working with several years ago. As I look at it, I think the democratization of AI is what has been unlocked. We put incredible power, in terms of prediction, into the hands of more people. And I think that will open up all kinds of opportunities to simplify and enhance customer experiences, making the lives of our associates easier. So I do think there’s a bit of a shift, and a lot of it has to do with accessibility.
Question 3: What are some of the early learnings in Generative AI for you?
First of all, if you talk to most people, I think the rate of change we’ve become accustomed to continues to compress. What we’ve seen over the last year, and even what we’ve seen with some of the recent announcements over the past couple of weeks, is that the acceleration of capabilities and the acceleration of benefits that you can get from this technology, is unprecedented.
So how you think about your business, how you think about managing it, and how you think about up leveling it has put a lot of pressure on existing structures. I think from my perspective, our goal has been to solve how you shift from a world in which the majority of your analytics are developed, supported, and maintained within an analytics organization, to one in which the analytics organization is part of a much bigger ecosystem in which businesses are using out-of-the-box capabilities with AI, third-party vendors are coming in, and the capability is being deployed far more broadly. So that’s a bit of a learning curve.
The good news is, if you have solid processes in place, and you have solid controls, then you have a basis around which you can start to build things out. I would imagine that if you don’t, you’re probably feeling rather exposed right now.
Today, everywhere we have a decision point being executed either by a person, through automation, or via AI (through prediction), we always talk about how to have the right control process and the right control framework. If it’s in a manufacturing facility you may have in-line quality checks. If it’s technology, we have deployment checks and monitoring around that. It’s no different, I think, with AI. You need to embed your solutions within a control framework. And that was probably a little bit easier to manage when everything was coming out of an analytics organization. Now you have to think about that at a different scale. So that’s definitely one of the learnings.
Another big one for us was engaging the business on the journey. To some extent, the exhilaration and excitement in the media around Generative AI has grabbed a lot of attention. And it’s grabbed a lot of attention, not just for Generative AI, but for how advanced analytics and data can have an impact and transform the business in general. So there’s been a real opportunity to have more discussions around the organization that enables us to work on some of the change aspects and educational awareness with business leaders.
The third learning is that you have to scale. It’s very hard to generate value through individual point solutions. But…if you build a capability in a way that’s beneficial on one decision point, and then another decision point, and another decision point…you can do that at scale. And that’s where you start to see some of the opportunities.
Some of the solutions out there are inherently built to scale, but I think a lot of the use cases basically amount to: solve this problem here, solve that problem there. But what you come to realize is that these solutions are just text summarization around a particular set of information that could potentially be used by a much broader range of employees. Ideally, you want to make that available so you can tailor it to individual populations. That’s where you can start to see some of the real benefits.
Bonus: This concept of ‘Buy versus Build’ is often a technology decision that you, as an IT leader, have to make. And with Generative AI, it feels like we’re at the beginning of a whole new set of technologies. What’s the ‘buy versus build’ thesis that you believe we should be thinking about?
I think it’s a little bit of a forcing function because of the speed and pace in the space today. The underlying fundamentals on this haven’t really changed. We should be building things in-house that are truly proprietary and differentiate us on a competitive level. That’s where I would prefer to put our assets and our professionals who are focused on building data and analytics products.
However, what I’m seeing is that the platforms are accelerating a lot of the capabilities that they’re offering from a Generative AI point of view, especially around commoditized capabilities, and we want to be able to leverage where they’ve already put in the work. So we’re additionally having discussions with our partners about what’s on their roadmap. For example, there have been a few cases where we saw tremendous value in leveraging Generative AI to better equip people to answer questions. However, after talking to our partners, we realized that was already on their roadmap. So we’ll apply our resources in other areas for the time being.
The third consideration around “buy vs. build” is that there are certain cases where we see unique emergent capabilities that we don’t currently possess the skill to advance. For example, startups that sit outside the large platform providers. Many startups have very unique capabilities on the outside that can accelerate a niche area where we want to drive value. So we’re working with emerging technologies as well.
I think because of the pace at which everything is moving, some things we thought were pretty novel ideas, we realized, were cases that just about every company wants. So we need to figure out just how much of the world is already on the large providers’ roadmaps. This enables us to assign our resources to other tasks that help us become more competitive.
Question 4: What are some of the issues, or maybe even gaps, in technology that you believe need to be addressed? I know this isn’t a technology-only issue. What are the headwinds that you have to address?
You have to think about scale to really understand the gaps. Point solutions are easy. You can go online and say: “Give me a recipe for a rustic Italian meal that is inspired by this part of the Italian coast.” Anyone can do that. It makes it seem very easy. It sets everyone’s expectations super high, before they start to see the gaps.
The real challenge is when you want to evolve it at scale in a way that can support your business. First, the tech stack starts to evolve. We went from a traditional tech stack that we had become fairly accustomed to, to a data and analytics tech stack. Now, we have to figure out how we want to incorporate it into both market-based and proprietary LLMs, and additionally fit everything into our architecture and the supporting set of capabilities around it. A lot is still being defined, it’s very new. The majority of companies, unless you’re actually a tech services provider building these capabilities, probably have a gap now because it’s novel.
For the other gaps, I think we need to take a hard look at changing the way we work. One of the discussions that we’ve been having with the business is: How are these capabilities going to change how we work? What are the skills and competencies that we need to start building within our organization? How do we help our people on that journey? Where are we incorporating new tools? Where are we working with our teams to grow?
You have to shepherd people along the path. Not because people don’t have the capability, but because it’s a chain. You need coding, but you also need prompt engineering. To some extent we now have solutions that we’ve deployed inside the business, and we’re working with people that are not technologists. So we have to ask: How do you leverage some of the Gen AI capabilities that we provided? How do you write prompts? How do you write good prompts? How do you evolve? How do you get the right information out? So initially it’s going to be about raising everyone’s skills and knowledge on the business side.
The final gap is going to be ensuring that this technology, like any manual or technical process that you’re building out, has the right, and most efficient, controls.
At the end of the day, the mindset needs to be less “I’m going to build an individual chair as a craftsman” and more “I need to build a factory that can produce furniture at scale.” The same thing is true here. I don’t need a use case. I need a system that can support multiple use cases in a way that is efficient, effective, reliable, and responsible.
Question 5: Responsible AI: How do you think about it? What does it mean to you?
It means a few different things. One of which I would say is that no matter what industry you’re in, you’re in the business of trust. You need to make sure that you’re thinking about responsibility and you’re thinking about how you’re using this technology in the right way.
Additionally, we want an environment where we’re bringing in and hiring people who can feel confident about the processes we’re going to exhibit and sustain, and the ethics around how we’re doing work. So we’ve put a lot of time and effort into refreshing some of the practices that we had around responsible AI. This includes enabling them to be ported across the broader organization, which brings us back to that concept of democratization. We have a framework in place that’s pretty focused on making sure that our AI is reliable and performative. It does what it’s supposed to. It’s compliant with privacy and trust. It’s secure. But also: It’s safe. We’re in the business of food. We want to ensure that our algorithms are performing in a way that accounts for the safety of our associates, our customers, and society at large. There needs to be accountability as we deploy these solutions, which goes along with transparency and explainability.
Finally, the other key aspect that we take a hard look at is around fairness. At the end of the day, our goal at 84 51º is to make people’s lives easier.
I’m very fortunate to be in an organization where we want to have a positive impact on the world. We want to make sure that what we’re doing from an analytics perspective reflects all those values, and all those ideals, both inside the company, but also with regards to the people we attract to come work here, who have similar standards and similar expectations.
Todd James is currently the Chief Data & Technology Officer at 84.51°, a retail data science, insights, and media company helping The Kroger Co., consumer packaged goods companies, agencies, publishers, and affiliated partners create more personalized and valuable experiences for shoppers across the path to purchase.
A driver of digital transformation, Todd spent 15 years at Fidelity Investments where he held a variety of key strategic leadership roles. An innovative leader, he built the global data and analytics organization for Fidelity’s Workplace Investing and Health Care business units. He also led efforts to modernize servicing and operations through applied artificial intelligence, and automation and directed Fidelity’s Cross-Enterprise AI Center of Excellence (COE).
Prior to Fidelity, Todd led a strategy consulting practice at Deloitte, directing strategic engagements with global Fortune 500 and government clients. As a director at Resource Consultants, Inc. he built and led a technology services business unit while also overseeing corporate IT. Prior to his business career, Todd was an officer in the U.S. Coast Guard where he held leadership roles in IT, information security, and shipboard operations.
Aside from a B.S. in Mathematics and Computer Science from the U.S. Coast Guard Academy, Todd also earned an MBA from The College of William and Mary, and a M.S. in Computer Science from the University of Illinois. He is an editorial board member for CDO Magazine.
