Formulas + runWhen
This guide shows how to spend credits where they matter: on the journalists most likely to care about your story.
- Formulas handle extraction, cleanup, scoring, and routing logic for 0 credits.
- Run When (displayed as
runWhenin the UI) is a conditional gate that skips expensive columns unless a row actually qualifies. Think of it as an "only run this if..." rule.
Why this matters
We see teams overspend by running AI on every row before filtering. That means credits burned on journalists who were never a good fit.
Use formulas first to triage and structure data, then run AI only on the journalists who qualify.
Fewer, Better Pitches
Treat formulas as your first-pass triage. The more journalists you filter out before AI runs, the more you spend on the ones who actually fit.
Cost comparison examples
| Workflow pattern | Rows | Cost per expensive run | Total credits |
|---|---|---|---|
| Run AI on every row | 1,000 | 1 credit | 1,000 |
| Formula triage + runWhen gates AI to top 30% | 1,000 | 1 credit | 300 |
| Formula-only transformations | 1,000 | 0 credits | 0 |
In this example, formula triage + runWhen saves 700 credits—and focuses your AI on the 300 journalists who actually qualify.
What you will learn
- Formulas: complete syntax, function reference, and PR workflow examples.
- runWhen: conditional execution for AI, enrichment, and agent columns.
- Cost Optimization: formulas, filters, and runWhen strategies to save credits.
Recommended rollout
- Add formula columns for cleanup and scoring (
Email Exists,Priority Tier,Needs Enrichment). - Add
runWhenconditions to expensive columns—gate AI to qualified journalists only. - Preview on sample rows, then run in bulk.
- Track credits saved per workflow and refine thresholds monthly. Small changes add up.