WWT CTO On Driving AI Wins, Nvidia And GenAI’s ‘Real Secret’

WWT CTO and AI expert Mike Taylor explains how the $20 billion ‘AI-first’ company is winning AI deals, creating new generative AI technology, and what a successful AI strategy looks like for a solution provider.

World Wide Technology’s longtime technology guru, Mike Taylor, is a driving force in the $20 billion company’s artificial intelligence strategy—from helping create new generative AI tech to successfully implementing AI solutions that drive actual business outcomes for WWT’s massive customer base.

“You shouldn’t be making a single technology investment today—whether it’s security-related, infrastructure or your cloud partnerships—that doesn’t contemplate existing or future use cases for AI,” said Taylor, WWT’s CTO and executive vice president of services.

St. Louis-based WWT has declared itself an “AI-first company” and is one of the top AI solution providers, having conducted a slew of successful AI in-production use cases while also building its own generative AI technology.

“AI-first means it needs to be present in every conversation, every topic, every domain expert needs to have AI relevance in terms of how they’re thinking about what they’re doing,” said Talyor.

“These AI initiatives and how they really take off, they have to be CEO-led. CEOs have to be in the room driving the discussions and driving the alignment between the business and technology teams. And without that, companies are going to be left on their back foot, or maybe worse, a few steps behind,” he said.

WWT’s AI Charge

WWT has been transforming itself over the past two years to become an AI leader.

The company has invested $500 million in several AI-related arenas, including building AI innovation centers where customers can play with technology; revamping its 10,000-plus workforce into AI experts; and creating homegrown GenAI products like Atom Ai and its Regulatory GPT Assistant.

“The real secret of generative AI is when you can wrap it around data sets or use cases. We’re seeing the first use case might take awhile because you’ve got to organize your data, you’ve got to build this thing out—but the trail of extended use cases and adoption of that is happening so much faster,” said Taylor.

In an interview with CRN, Taylor takes a deep dive explaining successful AI customer use cases, WWT’s tight partnership with Nvidia, how to build innovative generative AI technology and WWT’s overall AI strategy.

Explain to me one of WWT’s best AI use cases.

There’s a large mining company that we’ve been working with where we’ve been able to take machine learning techniques around health, safety and efficiency and add a generative AI component to it that now democratizes the health, safety, policies, etc. for the customer.

And [we’ve also added] the ability to interact with that data via language, like ask questions and have a conversation with some machine learning algorithms and outputs that we’ve had.

How did it start?

We started collecting telemetry off equipment that they use to run and operate their mines. Pulling that equipment in and then using machine learning algorithms to try to predict equipment, engine failure or other issues with these huge pieces of equipment. If one of them breaks down, it’s a significant impact to productivity. Their tires are like the size of a small home in some cases.

We used extraction methods to get this data off the devices themselves, get it into a data lake and a set of data lakes, where we then started doing machine learning to try to look at events ahead of an engine or an equipment failure that would help us better predict when something needed maintenance. That was where we started, which evolved to all kinds of crazy use cases.

So we created machine learning data lakes and all this stuff that we collected. Then we started putting language on top of that data to better help people interact, ask questions, observe patterns, maybe introduce other predictive things into that model that was going to make a model better. And that’s the real secret of generative AI.

It’s when you can wrap it around data sets or use cases. We’re seeing the first use case might take awhile because you have to organize your data and build this thing out. But the trail of extended use cases and adoption of that is happening so much faster.

So you go from this two-dimensional interface with an analytics platform that’s doing visualization to a language-based opportunity to have a conversation around the data you want to understand and how you want to examine it.

What is another good successful AI use case for WWT?

In the retail markets today, margins continue to become depressed against inflationary pressures—whether it’s goods that go into the production of those items that the retailers are selling, or in fast food the protein that goes into the meals that you and I eat.

All of that pressure is creating the desire and the will to do forecasting around consumption, labor, and how do you get supply and demand curves to meet better in terms of the retail market.

Where we participate in a lot of digital transformation is on the mobile app side, the customer-facing web and omni-channel stuff that we’ve done now, including generative AI, to help a broad distribution of retailers be able to ask questions of the data that help them better inform staffing and procurement practices around the goods that they need to sell.

Before, it would be somebody analyzing a spreadsheet, multiple tabs or having to click through filters, do other things. Now to be able to go in and ask questions and get responses back around peak times for consumption, or vulnerabilities they have in their staffing plans around where labor meets their customer in the physical retail stores—those are exciting examples.

What are two of your proudest AI products launched at WWT over the past 12 months?

The Atom Ai chatbot that we developed internally.

What we’ve been able to learn from the development of that, which has then produced speed to market with our customers, is emblematic of exactly what we want to continue to do over and over again.

When you get one of these Atom Ai [instances] set up, you’ve got your data structure set up, you’ve got the model fine-tuned, you’ve got your vectoring done in the right way—there’s now a trail of use cases that we can turn on like ‘that,’ which extend beyond supplying general knowledge.

So the spawn from Atom Ai is another one where we built a small- or medium-size language model that’s helping our regulated customers interpret emerging regulation and produce policy procedure documents that can help meet those regulatory requirements as they’re being imagined. This is our Regulatory GPT Assistant.

Financial services, utility and health-care customers that we’re working with, the amount of regulation that they’re under and how quickly that is being developed or being developed in retrospect is a lot.

This model that we’ve created is now a validated source for them to go seek out, ‘What are the regulatory compliance [issues] they may be subjected to? What are points and positions that we’ve developed with other customers around how they would represent their policies, procedures and compliance with those regulations?’

And where it took us a little while with Atom Ai around regulatory compliance, we were able to spin [Regulatory GPT Assistant] up in a matter of weeks and start getting value both for us and for our customers in a really short period of time because of the work we had done ahead of it.

What is your advice to companies wanting to drive AI traction with customers and/or internally?

These AI initiatives and how they really take off, they have to be CEO-led.

CEOs have to be in the room driving the discussions and driving the alignment between the business and technology teams. And without that, companies are going to be left on their back foot, or maybe worse, a few steps behind. That’s something we talked to our customers about.

[WWT CEO] Jim Kavanaugh is a really good example of somebody who is a lifelong learner.

He stepped in, recognized this, and has been driving the organization for over a year now on, ‘What is it going to mean to our business? If it means this to our business, what does it mean to the market more broadly? Who are the partners?’ But it’s CEO-led.

It has got to be that CEO working with their executive teams to harness the value and the opportunity here.

How did you get such a deep partnership with AI superstar Nvidia?

We’ve been doing this work before it was cool. We were doing it because it was delivering a lot of business value, in particular in the machine learning context: reinforcement learning, machine learning—these are optimizer models that had been created.

These were things we’ve been doing for a very long time, not because Nvidia was cool or because AI was the rage.

For us, it’s just been part of our DNA, where we knew that AI was a really impactful tool that sat very closely and adjacent to security and digital outcomes that we’re working on with our customers.

Ten years ago, we brought in a very tight cohort of data scientists and now have gone and built that out to now 400-plus people who are focused on data science from a modeling standpoint and from a training perspective.

And now with retrieval augmentation and vector databases and all these other things that are now cool and in vogue , we’re like, ‘Yeah, let’s accelerate because more of the markets are starting to get it. But let’s also understand we’re not some fly-by-night company that’s just inventing it as we’re going along. We’ve got a long history here.’ And we’re proud of that.

Is WWT an AI-first company?

One hundred percent. If you are a business leader or a technologist, you need to be drawing the corollary between your traditional practice [and AI].

Let’s just say you lead network for a large organization or you’re a line-of-business leader in supply chain or sales operations—you have to become literate on what the possibilities of AI are. Also, you need to understand what are the functions within your business that AI and machine learning, reinforcement learning, as well as generative AI to this whole mix here of, ‘How can those technologies impact not just the speed of my service, the quality of my service, but where does it help me invent or think about evolving what you do today to hire orders of service or new service offerings for your customers that you haven’t considered before?’

You shouldn’t be making a single technology investment today—whether it’s security-related, infrastructure, your cloud partnerships—that doesn’t contemplate existing or future use cases for AI.

So AI-first means it needs to be present in every conversation, every topic, every domain expert needs to have AI relevance in terms of how they’re thinking about what they’re doing.