
Andrew Hanna

Andrew Hanna

The argument in three sentences: most Salesforce changes are not made by developers, they are made by admins and consultants who have never opened a terminal. Salesforce Ben's 2026 survey analysis puts change sets at 41.8% as the primary deployment method for admins, against 2.6% for DevOps Center. Any DevOps tool whose first screen assumes Git fluency has quietly excluded most of the people who ship on this platform.
The role split in the data is stark. According to Salesforce Ben's analysis of its 2026 developer, architect and admin surveys:
Read that again from the tooling side. Developers and architects have largely left change sets behind. Admins have not, and they are the largest population making changes. The gap is not appetite. It is that the tools built for the first two groups were never usable by the third.
Because the Salesforce customer base does not look like a software company. A typical team is two admins, a consultant on retainer, and no dedicated DevOps engineer at all. When a tool's happy path starts with cloning a repository, choosing a branching strategy and resolving a merge conflict in a diff viewer, the honest translation is: hire someone else first.
That is why the change-set number refuses to fall. It is not that admins prefer a tool with no history, no rollback and no way to move metadata between unrelated orgs. It is that change sets are the only option that never asked them to become a developer on the way in.
The blunt version: a DevOps product that requires Git fluency is not a Salesforce DevOps product. It is a developer tool that happens to deploy Salesforce metadata.
Not a simplified developer tool. A different entry point onto the same engine. Four design tests separate the two:
Git still does the work underneath all four. It just stops being the interface, so developers keep their repository, branches and CLI while admins work on the same history from a UI.
It is moving in that direction, which is the strongest evidence the reframe is real. The next-generation DevOps Center is native rather than a managed package, enabled in Setup, and built around work items, visual pipelines, per-item promotion and a DX Inspector that tracks org changes, with agentic pipeline management through the DX MCP server (Salesforce Ben).
Salesforce moving DevOps into the core platform is a category signal: the platform owner now agrees that release management belongs to the whole team. It also raises the bar for everyone else. If your differentiation was "easier than the CLI", the free native option is coming for that.
You are the person who inherits the consequences of the split. Three practical moves:
This is the assumption Serpent was built against: ticket-based deployments with Git in the background, one-click rollback and AI code review on every plan, including the free one, so the admin and the developer ship through the same pipeline.
Do admins really need DevOps, or just better change sets?
They need what DevOps provides: history, review and undo. Change sets structurally cannot give them those, since a change set keeps no version history and cannot be replayed.
Does a no-Git workflow mean no source control?
No. It means the commits are generated for you. The repository still exists, and your developers can work in it directly.
Will the native DevOps Center make third-party tools unnecessary?
For simple org-to-org pipelines it will cover more teams than the managed package did. Backup, rollback, package and ISV workflows and automated testing are still where the gap sits.
Where should a team with no DevOps engineer start?
One pipeline, one production org, one ticket per change, one approval. Push a single low-risk release through it end to end before migrating anything else.
Commitment free!