Beyond replacement: How AI is enhancing PaaS offerings

AI and PaaS

In a BrightTALK Cloud Cover webinar, "The Growth of Platform-as-a-Service (PaaS)," a panel of experts in the PaaS market talked about their predictions for the cloud model. Unsurprisingly, a big topic of conversation was AI and its connection to PaaS.

Angela Andrews, senior solution architect at Red Hat, shared that AI is at the forefront of consumers' minds across the board.

"Everyone's talking about it. It is top of mind for businesses and C-suite individuals, and we're seeing it," she said. "We're already seeing the cloud vendors providing AI PaaS solutions in their portfolios, and more and more companies are getting into the space."

What exactly are the features these professionals are pursuing?

AI PaaS features

Here are some examples of AI features within PaaS services:

• Retrieval-augmented generation. RAG assists in connecting and cross-referencing information to provide LLMs with up-to-date information. Consider chatbots or search engines. These tools use RAG capabilities to provide end users with relevant information in an instant. This could optimize all kinds of businesses, and it relies on the combination of AI and PaaS technologies.
• Automated machine learning. AutoML helps users create ML models by automating tasks to simplify the process. By taking away most of the complexity of building, such as data processing and algorithm selection, the process of building models becomes more user-friendly.
• Cloud-native tools. Cloud-native applications are built with the knowledge that engineers will run and host them in the cloud. Several AI tools for cloud-native applications can help with traffic monitoring, container management, orchestration and storage.

These are only some of the tools that AI can provide engineers and users through PaaS offerings. But what are the benefits of moving to AI PaaS services?

Benefits of AI in PaaS

Dave Linthicum, founder and lead researcher at Linthicum Research, summarized the benefits of AI in PaaS: It's the easy button.

People don't want to make these complex engineering decisions. They want to make decisions around how the business is going to move. They need the easy button.

"People don't want to make these complex engineering decisions. They want to make decisions around how the business is going to move. They need the easy button," he said. "PaaS is able to do that because you're able to deal with known resources and with AI using known integration patterns."

The panelists agreed that PaaS technology exists to "abstract away the complexity" of application engineering. Adding AI to PaaS services optimizes what PaaS can do. Below are some of the benefits and optimizations afforded by AI in PaaS.

Automation

According to Linthicum, the core benefit is task automation. With these tools, users can access and deploy prebuilt AI models to assist in architecting applications. Users can also automate data processing to provide their organizations with real-time analytics that inform important decisions. For security, AI automates monitoring and can conform to specific compliance needs.

"The ability to abstract the user, the consumer, of the AI systems away from making many engineering mistakes is going to be the whole core benefit," Linthicum said. "This is about automation."

Cost efficiency

By design, PaaS saves customers money as CSPs provide them with compute capabilities without the need to maintain physical hardware. This is already a cost-efficient way for many companies to operate. Now, businesses can take advantage of AI architecture through a similar pay-as-you-go model or implement AI tools within their existing subscriptions or licenses.

Because users can access AI tools at a lower cost threshold, they can also benefit from faster time to market. This boost in production can help organizations generate revenue faster, making AI PaaS services a viable choice for cost efficiency.

Ubiquity

The recent ubiquity of AI features requires relevant stakeholders to become familiar with this new technology. The three major CSPs, Amazon, Microsoft and Google, each have their own AI PaaS services -- AWS AI Services, Azure AI and Google Cloud AI. Because these three companies have a collective market share that represents a majority of cloud users, many enterprises have access to what might become industry-standard AI services.

These services support existing workloads and resources, meaning simpler deployment and integration. Users and engineers of all skill levels can use AI technology and add valuable skill sets to their toolkits -- and resumes.

Partners, not competitors

Given what we know about how AI can optimize PaaS offerings, it's justifiable to wonder whether AI technology will, or can, replace PaaS entirely. But the webinar panel was far more skeptical.

Jessica Orozco, senior vice president of global sales for Platform.sh, gave a firm answer as to why this is the case.

"A lot of AI co-generation is heavily subsidized right now," she said. "There are a lot of big VC firms behind it all. I think once you have that subside, there's going to be some real economies of scale to take hold and get to a place where it's more cost-effective. It always comes down to cost, right?"

Linthicum echoed Orozco's thoughts, saying there's "an impedance mismatch" between tech companies and their enterprise customers.

"[Private equity money] is kind of changing the priorities, and probably obscuring them a bit, because I think we're not focused on what enterprises need," Linthicum said. "We're focused on what the private equity people want small tech companies to build. And I think those are two different problems."

While 85% of businesses are currently using AI services or tools, this is because of the surge of activity and innovation within the AI sphere. Once this phenomenon starts to stagnate, and organizations notice dissonance in what they need versus what providers offer, AI use could begin to settle. There's also the continuous need for upskilling as companies innovate on their AI technology.

This is where Andrews sees PaaS as being an answer to the complexities -- both financial and technical -- brought by AI.

"We continue talking about how the complexity is only going to grow with GenAI and other AI applications, and I think it's going to be a nice little kind of marriage where AI comes together with PaaS to remove those complexities," Andrews said. "I think PaaS is just going to be the natural choice that most companies are going to be able to make moving forward."

Everett Bishop is assistant site editor for SearchCloudComputing at Informa TechTarget. He graduated from the University of New Haven in 2019.