Click here to enjoy the entire Amir Sharif interview at

Disruptions and restrictions caused by COVID-19 prompted many SaaS enterprises to speed up their cloud adoption as part of their streamlined digital transformation initiatives.

However, while moving to the cloud is becoming a permanent trend, especially among SaaS companies, a gap is materializing between management and DevOps. These two entities often struggle to get to a consensus on cloud adoption strategy and implementation. Amir Sharif, our VP for Product and Marketing for Opsani, dug deeper into this conflict in a recent interview with Digital Anarchist. 

Both management and DevOps agree that cloud adoption is now essential for SaaS companies’ growth: “There is a faster march towards cloud and cloud-native technologies,  leading to faster releases and reliance on agile technologies and faster delivery to market,” Amir pointed out.

Amir described SaaS companies as “the vanguard of cloud adoption and cloud-native technologies.” As enterprises move to the cloud, he said, they also increased their spending containers: “People are adopting containers by large numbers. 60% of applications are rolling on containers. At least 50% of workloads are orchestrated by Kubernetes.”

The First Gap: Budget vs. Efficiency

However, as Amir unpacked, management and DevOps have differing views on how they should proceed with cloud adoption. Management wants to use cloud while sticking to a certain budget. Developers tend to be headstrong and bullish when it boils down to their work, which is why many go for multiple cloud adoption.

While partnering with a single cloud vendor suits any corporate strategy, developers tend to want to have the technologies that simplify and streamline their work.

“It comes down to corporate strategy and corporate negotiation and whether a single vendor can give you the best offer. But, developers, as a whole, like to do what they think is best of them. Despite top-level edicts, they will do what allows them to develop code faster and go for the best of breed technologies.”

Amir encourages SaaS companies to give developers the freedom to pick the technology they think is best for them and not focus on cost so much.

The Second Gap: Level of Quality

SaaS companies release 70-80% of their code and applications daily. In terms of the quality of that code, management and DevOps don’t always see eye to eye.

“Management believes that the code being put out is of utmost quality. When you ask the DevOps people ‘what really drives you?’, the answer is staying within a certain time and making sure that is good enough.”

Amir recognizes that SaaS enterprises are always racing to get the new and improved features out. Developers are content to deliver code and features that work as long as they meet requirements and nobody complains.

“You don’t put a lot of thought into crafting quality, making sure it’s the most efficient code, despite what management thinks,” Amir said. “That’s the whole premise of cloud-native. You can decouple workloads as long as the API contract is honored; you can release various parts of the code very quickly.”

“The reason why that’s possible is because functionality is now modularized. You can improve a part of the code without having to recompile the whole shebang.”

The Role of AI in Cloud Management

IT environments are getting more complex and more challenging for people to manage. Multiple clouds create diverse sets of infrastructures. Monitoring and managing these environments and making sure they perform at an optimal pace is no longer within human capability.

According to our recent poll on cloud adoption, 91% of 1000 C-level executives at SaaS companies said they were “highly confident” or “confident” that their cloud applications were running efficiently. To be frank, their confidence is probably misplaced. 

Most enterprises are overprovisioning resources to keep their cloud operations at an optimal level. On top of that, cloud monitoring and resource provisioning are not automated. Both result in sub-optimal performance and higher cloud costs. 

“Is there complexity? The answer is yes, absolutely,” Amir affirmed. “If you try to (manually) optimize your workload, it’s going to be extremely difficult.”

Finding the right configuration to optimize your cloud infrastructure, jobs, processes, and workloads is simply beyond human scale. Cloud cost optimization needs to be fully automated and AI-powered.

“There’s a whitepaper that we have at that talks about the canonical example for a cloud-native application, which is the Google Online Boutique. If you try to optimize the entire application, you need to consider 75 quintillion different permutations,” Amir said. 

Going through each possible configuration, testing, and analyzing to determine the best for your cloud infrastructure is not just impractical. It’s impossible.

“That happens to more than the grains of sand on earth! You need to have tools to help you cope with that complexity. That’s why you’re seeing adoption of AI.”

Click here to enjoy the entire Amir Sharif interview at