Why Cloud Is the Go-To Data Platform for AI

Why Businesses Harness AI in the Cloud

Many businesses are still trying to harness the power of GenAI. Early adopters have moved beyond experimentation to building new applications and services that improve customer experiences and enable employees to work faster, smarter and more efficiently. As they do, many rely on the cloud to drive innovation.

The major cloud providers — AWS, Microsoft Azure and Google Cloud Platform — deliver the infrastructure, software tools and foundation models companies need to develop and scale their GenAI applications, says Futurum Group analyst Bradley Shimmin.

“The major cloud platforms wrap it all in a cohesive package that makes it easier for companies to provision the resources they need to build generative AI applications,” he says.

Enterprises today have a massive skills gap when it comes to keeping pace with the rate of change in the modern software stack, which includes AI, he says. It’s much easier to turn to the cloud.

“It’s all about time to market and simplicity,” Shimmin says. “Companies can very well build all of this out on their own, but they’re choosing to partner with these cloud providers because it derisks their investments. Not only does it speed time to value but they don’t have to worry about compliance for security or key management because the hyperscalers have already taken care of that.” 

PGA TOUR Enhances the Fan Experience With AI

The PGA TOUR began exploring GenAI two years ago, collaborating with AWS as the cloud provider developed its AI tools.

In addition to TOURCAST, the TOUR has created an internal GenAI tool to support employee productivity and is exploring the development of a GenAI -powered analysis tool to help professional golfers analyze their play. The TOUR is also assessing the opportunity to develop a tool that could provide broadcast teams instant talking points after every shot, Gutterman says. 

The organization uses AWS’ Amazon Bedrock, which enables organizations to build and deploy GenAI tools. The PGA TOUR built its GenAI tools in its own secure environment within AWS. The TOUR uses its own proprietary data through retrieval augmented generation and choose the best large language model for its needs. It feeds its internal data into Anthropic’s Claude LLM to generate customized commentary.

The TOUR, which began using AWS in 2012, runs its entire tech stack on the cloud platform. The TOUR’s ShotLink technology, powered by CDW, has tracked every shot since 2001, and the data is housed on AWS alongside the organization’s media archives containing all tournament videos.

“It’s our own cloud environment. Nobody else has access to it, so our data and content assets are protected,” Gutterman says.

To ensure responsible use of AI, the TOUR uses the Bedrock Guardrail tool to help eliminate toxicity and bias, ensure statistical accuracy, and maintain brand alignment, he says. Building effective AI commentary requires sophisticated prompt engineering, which Gutterman calls “context services.”

For example, LLMs understand general golf concepts, but need guidance on specific nuances of the sport. “You have to tell it things like, ‘Don’t pay attention to the random golf data in your model,’” he says. “These models have gotten data from everywhere. It could pull context from all kinds of things if you’re not engineering the prompts correctly.”

To build its GenAI tools, the PGA TOUR’s product managers and data science and scoring groups have collaborated with experts at the AWS Generative AI Innovation Center.

The organization’s internal chatbot allows staff to query 30 years of media guides stored as PDFs. This helps staff quickly research historical content, Gutterman says.

The TOUR launched AI match recaps for the 2024 Presidents Cup last fall, then added real-time, shot-by-shot commentary this spring to TOURCAST at the 2025 PLAYERS Championship. Both are popular with fans.

“It’s offering greater context around every shot a player hits, and enabling our fans to personalize the way they consume golf,” he says.

DIG DEEPER: How to train your artificial intelligence bot with chain-of-thought prompting.

Alaska Airlines Reinvents Travel Search

In Seattle, Alaska Airlines has reimagined how customers can plan their vacations. Instead of juggling dozens of web browser tabs, travelers can now use a GenAI search tool that delivers destination recommendations and allows them to view photos, videos, route maps and book flights in one place.

If customers want ideas for getaways, they can use natural language to ask questions like, “I want to take a budget-friendly trip to the beach this summer,” and the tool will make suggestions with flight options.

Called Trip Inspiration With AI, the tool is built on Google Cloud, integrating the airline’s operations and customer data with Google tools and different LLMs, says Natalie Bowman, Alaska Airlines’ vice president of digital experience.

Click the banner below to learn data governance strategies that help with artificial intelligence initiatives.

“We wanted to play in the AI space but do it in a way that drives our business. So, we looked at our customer pain points,” she says. “Flight search is, in some ways, a suboptimal experience. We require guests to give us specific information when they’re still exploring. It felt like the right way to use AI: to allow people to tell you what they want.”

The project was made possible because the airline moved about 85% of its guest-facing platforms to the cloud over a six-year modernization effort completed in 2023. Alaska Airlines uses Microsoft Azure for most customer and operations data and Google Cloud for its Google Flights QPX shopping engine and its BigQuery data warehouse, which powers AI models for customer insights and marketing.

The airline’s principal product manager and data scientists built the tool with Google engineers, using Vertex AI, Google’s software for building and deploying GenAI apps, along with BigQuery, QPX, Google Gemini and Google Maps.

Alaska Airlines built a second version of the tool using OpenAI’s ChatGPT that focuses on search results and includes global airline partners, while the Google version is more feature-rich, showing map routes and destination details such as museums, Bowman says. Right now, the airline chooses the version customers see. 

The airline built guardrails using a custom tool called Quality Assurance Response Liaison to prevent inappropriate requests. With safeguards in place, the company rolled out the current version of the tool in October 2024. It generates the same amount of revenue as a fare sale when promoted on the home page, she says.

“It means we can drive revenue without discounting fares, so we keep our margins high,” Bowman says. “It’s a great business result.”

UP NEXT: Businesses must personalize their AI to remain competitive.

Nasdaq Transforms Trading With GenAI

Nasdaq has built cloud-based GenAI tools to power some trading operations and improve employee productivity. To launch GenAI initiatives, the global electronic marketplace leveraged its 15 years of cloud expertise using AWS, says Brenda Hoffman, executive vice president and CTO for Nasdaq’s Market Services and Financial Technology divisions. 

“Our organization had built the muscle around using cloud and how to have data securely in the cloud,” Hoffman says. “We had done that work for our existing products and services, so we were advanced.”

In 2024, Nasdaq developed and deployed several GenAI -powered products, including the first AI-powered stock trading order type approved by the U.S. Securities and Exchange Commission. Dynamic Midpoint Extended Live Order (M-ELO) uses GenAI to analyze 140 market factors, then automatically adjusts trade execution timing based on stock volatility, resulting in 20% more orders filled for institutional investors.

Three years ago, Nasdaq began exploring GenAI through proofs of concept for its products and across different internal disciplines.

When Nasdaq adopts technology, it does so deliberatively, deploying it securely with governance, oversight and responsible AI at the forefront, she says. The company has completed about 100 proofs of concept, building a governance structure with mandatory processes and feedback loops for each one.

“The big thing inside Nasdaq is a reflective review: What did we learn? What did we like? What didn’t work? Did it fail or pass?” Hoffman says. “We share that learning, so the next proof of concept takes advantage of everything we’ve learned.”

The oversight allows them to “go faster and do better with each revision,” she says.

Nasdaq has broadly deployed internal GenAI tools across different departments. The company uploads documents, so employees can query them and quickly get answers. Initially, people feared this would replace workers, but they’ve realized that it’s a productivity tool that reduces tedious work, not a replacement tool. 

“We are empowering our people to do more than they normally could do and be more productive, more accurate and faster,” Hoffman says.