Making Artificial Intelligence Work at WWT
In this blog
Real AI impact starts at home
When generative AI (GenAI) burst onto the scene with the public release of ChatGPT in late 2023, organizations of all shapes and sizes were eager to put it into action. This was certainly the case at WWT. From the outset of GenAI becoming a household name, we knew that in order to do it right, we would need some guardrails. In short order, we created our IT AI Center of Excellence.
The purpose of our team, which sits within WWT IT, does not revolve around building custom models or launching cutting-edge projects for experimentation's sake. We're about identifying real business objectives and helping teams solve them. Sometimes, this involves using the GenAI tools we already have. Other times, it requires homegrown development or integration of an off-the-shelf product. Whatever path solving takes, our goal is simple: practical AI adoption at scale.
Start with the problem, not the platform
When a team reaches out, it's usually not with a problem they want to solve but with a solution they already have in mind. They'll say, "We need to train a model" or "We want to build an AI agent."
Our first job is to pause and ask: But what are you actually trying to do?
In most cases, we don't need to build anything. The solution already exists in a platform like Microsoft Copilot or Atom Ai, our internally developed AI assistant. The real issue is awareness. People don't always know what tools are available or how to use them effectively.
That's why our team starts with education.
Demystifying the process
There's a lot of noise in the GenAI space, and buzzwords like "modeling training" or "building agents" can create confusion and inflate expectations. GenAI can be seen as a powerful yet wholly mysterious technology. Because of that, people think solutions must be incredibly complex. But what we've learned is that most of the time, a solution is much simpler than one would think.
For example, a team in human resources recently asked us to build an agent to simulate one-on-one conversations using our Integrated Management and Leadership (IML) framework. The goal was to help managers practice giving feedback and having tough conversations.
But when we looked into it, we realized they didn't need to build an agent. With the right content and a well-crafted prompt, Atom Ai could already do what they wanted. What we did have to do, however, was make sure Atom had access to IML materials in a sustainable, curated way.
To meet the human resource team's goal, we needed to devise a strategy to identify the right content, determine the best way to share it with Atom and establish a repeatable process to keep it fresh.
Read how WWT's Integrated Management and Leadership program is key to WWT's success.
Success with GenAI often requires us to demystify the technology. When we do, simple, elegant and business-impacting solutions can come to the surface.
Getting involved before things go off track
Not every GenAI project starts with us. For example, one supply chain initiative began as a proof of concept with a third-party vendor. The idea was to use AI to automate label scanning in our warehouses. Unfortunately, the company they selected (based on a trade show demo) folded mid-project, and the effort stalled.
That's when we got the call to step in.
We assigned one of our experts, Harry Kabbay, Lead Machine Learning Engineer, to embed with the supply chain team. He assessed the situation, worked with business and IT stakeholders, and helped build a better path forward using the tools and knowledge we already had.
Streamlining warehouse operations with AI
It was a multi-month effort, but it turned a failed initiative into a functioning solution. And it showed why this kind of internal AI partnership matters, especially before big investments are made.
We meet teams where they are
Our work spans a wide range of knowledge levels. Some teams are just beginning to explore prompt design, while others are experimenting with advanced concepts like agent-to-agent orchestration.
We don't push a one-size-fits-all model. We tailor how we engage based on where the team is, what they want to achieve and how ready they are to move.
That could mean a one-hour working session to clarify a prompt. Or it could mean assigning someone full-time to partner with a business unit for months. This variance influences how we're structured.
Some developers focus on platform stability. Others focus on enablement, spending time with teams to understand problems, guide strategy and translate between business needs and AI capabilities.
Learning as we go
When our team was first stood up, the environment was chaotic internally and externally. Everyone, including us, was trying to figure out what to do with this new technology. Early on, we assumed our work would be mostly solution-building. What we quickly learned was that education is just as critical.
And it's not easy to scale. WWT has thousands of employees, all at different points in their GenAI journey. Some need 101-level introductions, and others need help exploring highly specialized use cases. We're still figuring out the best way to serve that spectrum.
We also know we can't do this alone, and we try to work in parallel with WWT teams that are building GenAI solutions for our clients. Better coordination and shared visibility are helping all of us work smarter, and we are partnering more and more across those boundaries to align on our GenAI strategy, security measures and approved tools.
Outcomes matter
Our leadership expects more than activity. They expect outcomes.
The question isn't whether we're using AI; it's whether AI is making a difference. What have we improved? What decisions are easier? How much time have we freed up? These are hard things to measure, but they're the right things to aim for.
We've moved past the phase of GenAI for GenAI's sake. Our focus now is on sustaining value. That means delivering faster, working smarter and helping teams get to what matters without overcomplicating the process.
Learn how WWT defines practical AI
AI is how we work now
This isn't a temporary push. GenAI is becoming foundational to how we operate. Our team exists to help make that shift through education, collaboration and shared problem-solving.
We're not here to block or slow things down. We're here to guide teams toward scalable solutions that solve real organizational challenges and help WWT work smarter every day.
The work is messy, fast-moving and sometimes hard to measure. And that's a good thing. We're always learning, always improving.
For anyone trying to bring GenAI into their organization, remember that it's not about having all the answers. It's about asking the right questions and having the right people at the table to figure it out.