Delivering software into on-premises enterprise environments is a challenge many vendors are still struggling to solve. As IT leaders look for seamless distribution, effortless operations, and robust customer experiences, they quickly discover that most solutions in the market are limited to one-off packaging or manual deployment tooling. Tools like Replicated help wrap your Kubernetes manifests with licensing and delivery metadata, allowing customers to manually install and update their apps — but this manual process leaves too many operational gaps for both vendors and end-users.​

When teams talk about Bring Your Own Cloud (BYOC) enablement, the conversation often gets conflated with tools that only address small slices of the problem. At first glance, platforms like Nuon might appear to help you “run in your customer’s environment,” but what they actually provide is far narrower. Nuon simplifies one operational leg — orchestrating Terraform and Helm charts within customer environments via their agent-based mechanism. That agent retrieves and deploys your configurations locally, which can be useful for a very specific deployment flow. But that’s where their focus stops and even thats limited to specific cloud environments (no GCP, OnPrem or OCI support)

In the software industry, entrepreneurs have to go through 4 phases of innovation to build a product:

Idea: What change are they trying to make?

Design: How will their users interface with their product and services?

App: Build the core software using LLMs, and other building blocks

Product: Build the distribution engine to distribute their App to rest of the world

For the last mile to go from (Agentic) App to Product, organizations have to build multiple software distribution channels. The most common ones are OnPrem, BYOC, PaaS, SaaS and Agent-as-a-Service (AaaS).

Now in Part 3, we turn to the solution: What if you didn’t have to abandon your existing stack, but could augment it with a purpose-built control plane, one that handled tenant lifecycle, deployment orchestration, and multi-cloud complexity out of the box?

Imagine a data analytics SaaS startup, “DataCo”, that wants to build Redis SaaS as an example. They use Argo CD to deploy their analytics application (a set of microservices + some Spark jobs) onto the EKS cluster.

Modern platform engineering teams often assemble “DIY” control planes for their Products using open-source building blocks: Kubernetes for the runtime, Terraform for infrastructure provisioning, and Argo CD for continuous deployment. This stack promises cloud-agnostic flexibility and full control. But as many CTOs and platform VPs have learned, stitching these tools together into a robust control plane is hard.

On March 7th, 2025, we kicked off the inaugural session of SaaS Mondays, Live on Fridays, a new biweekly LinkedIn Live series dedicated to fostering discussions among SaaS builders, founders, and industry leaders. Hosted by Michael Cooper (GTM Lead, Omnistrate), the session featured insights from:

• Ermin Dzinic (AWS SaaS Specialist)
• Bill Tarr (AWS SaaS Factory)
• Kamal Gupta (CEO & Co-Founder, Omnistrate)

Our first discussion centered on the pressing question: How will AI impact SaaS builders? Below, we explore the key takeaways from this dynamic conversation.