Sub-Millisecond Workflow Steps
Most workflow automation is measured in seconds: a step that sends an email might take a second or two. That's fine for most scenarios. But some automation scenarios require much faster execution — high-frequency event processing, real-time data transformation, automation that runs on every keystroke or every record change in a high-volume environment.
Sub-millisecond workflow steps make those scenarios possible.
What Changed
Workflow step execution has two components: evaluating the step's logic (conditions, field mappings, branching) and executing the step's action (updating a field, calling an API, triggering another step).
The logic evaluation — the "should this step run and with what inputs?" part — is now optimized to run in under a millisecond for steps that don't require external calls. Definition lookups that previously involved database queries are now served from a compiled in-memory representation. Condition evaluation that previously parsed template expressions on each run now uses a precompiled evaluation plan.
Steps that don't have external dependencies (pure data transformation, field calculation, branching logic) run in microseconds. Steps that call external APIs or perform database writes are limited by those operations, but the surrounding overhead is minimized.
What This Enables
High-frequency event processing. A workflow triggered on every record update in a high-volume table can run without becoming a processing bottleneck. Thousands of workflow executions per minute are handled without queue backup.
Real-time computed fields. Fields that are computed by workflow steps — running calculations, formatting values, aggregating from related records — update in near-real-time when their dependencies change.
Responsive automation. In scenarios where users expect immediate feedback from automated steps (status auto-update, sequential field auto-fill), the response is now fast enough to feel instant.
Sub-millisecond steps open the door to automation patterns that weren't previously practical. The platform's automation capabilities just got more capable.