James Rose, chief technology officer at Morningstar, says 90% of financial services companies' cloud spending is on Amazon Web Services.
Microsoft is a distant second with about 7%, with Google Cloud accounting for the rest. But when it comes to artificial intelligence products and services, Morningstar is betting big on Microsoft's OpenAI.
“We didn't double down on the major cloud providers,” Rhodes says. “Microsoft offers out-of-the-box services today so you can get started right away.”
Big cloud providers are also in the business of providing AI products to companies like Morningstar, so should companies stick with their own cloud provider when building an AI ecosystem, or choose another partner or even An important decision emerges: whether to collaborate with multiple partners. .
“I think at the high end of the market it's going to be multi-cloud, and that's the reality,” says Spencer Kimball, co-founder and CEO of database startup Cockroach Labs. “So why would they change it for AI? I think they would embrace it even more.”
In deciding on OpenAI, Morningstar evaluated the terms of use of products available on the market because we were concerned about how the data we sent to providers would be used. We felt that Microsoft was more transparent than other companies about what they do with data and how it feeds back into their models.
Microsoft has made it clear to Morningstar that its data is only captured for debugging purposes and will be deleted after 30 days. That gave Rose peace of mind.
“I think what actually happened was that Microsoft was pretty advanced in enterprise use cases,” Rhodes says. “Honestly, it felt like Amazon and Google were asleep in the driver’s seat.”
Last year, when Morningstar introduced its AI chatbot, Mo, it combined its investment research library with Microsoft Azure's OpenAI service.
Cloud computing and enterprise software provider Appian has taken the exact opposite approach from Morningstar. AWS is the company's primary cloud provider, and the two companies also collaborate closely on his AI. “We already had a great relationship and we’re excited to build on that,” says founder and CEO Matt Calkins.
Appian announced in April that it had signed a deal with AWS to support its new private AI tools. Appian is an advocate of private AI, which he says represents three core values: Models must be trained using only company-proprietary data, data remains firmly under the control of the organization, and AI models are proprietary and never shared.
The partnership with Amazon Bedrock allows Appian to host large language models within specific compliance boundaries set by customers and privately customize those models. Cloud-based machine learning platform Amazon SageMaker allows Appian customers to create, train, and fine-tune their AI models using only their own data.
“It’s easier to stay within one ecosystem,” Calkins says. He credits his previous relationship with AWS with giving him an edge in AI. “We were probably going to pick them unless they messed up. We have a great history with them.”
Kimball said cloud providers, at least for now, have very different offerings when it comes to AI. “Clouds vary a lot from cloud to cloud,” he says. “There is no convergence yet.”
Large enterprises will likely be juggling several different AI products that reflect their approach to the cloud. But for small businesses, consolidating everything into one cloud and adding AI products and services could be more affordable and easier to work within the one ecosystem. there is. “I think the arguments around simplicity and discounts may be more powerful for smaller companies,” Kimball said.
Unstructurald CEO and co-founder Brian Raymond recently raised $40 million from Nvidia, IBM, Menlo Ventures, and more. The startup, which is at the beginning of the AI data journey, ingests unstructured data such as emails, documents, images, and videos and preprocesses it into a ready-to-use format with underlying models.
Unlike the cloud, which in the early days could be expensive to switch providers, moving between AI companies is easy and cheap. “It’s like fast fashion,” Raymond says. “Switching costs for any model are so low that few vendors lock them in.”
Raymond sees two different AI strategies emerging. There are vertical AIs that focus on applications that serve specific industries, such as healthcare or travel. and horizontal AI that can enhance processes like cybersecurity and customer support in any field.
“These cloud service providers are pursuing both vertical and horizontal strategies simultaneously to reach different market segments,” Raymond says.
ACI Worldwide has adopted a hybrid cloud strategy due to the nature of the payment systems company's business. The company has generative AI production products running on Microsoft Azure and AWS, but Google is still in its early stages.
Chief Technology Officer Abe Kuruvilla said the company recently deployed a chatbot powered by Microsoft's AI, in part because it was easy to deploy, but also because it complemented ACI's existing He said it was also because it was part of the license. “It’s pretty cost-effective,” Kuruvilla says. “We find that each of these providers offers add-on services at an incremental cost.”
But he is careful about overspending. Kuruvilla says companies need to closely monitor some of his AI spending to ensure return on investment and productivity gains are clear. ACI is currently testing Copilot for Microsoft 365 and expects to make a decision by June. However, the cost per user will be a factor in deciding where to deploy it.
“We're piloting it. Hundreds of people are using it and we continue to receive feedback,” Kuruvilla says. “More importantly, we are monitoring usage.”
