> For the complete user guide index, see [llms.txt](https://portraitresearch.com/documentation/llms.txt).

# Run and schedule screens

[Canonical page](https://portraitresearch.com/documentation/screening/run-and-schedule-screens/)

## Applicability

- **Task:** Run, schedule, inspect, and stop a saved screen
- **Access:** Portrait account; edit access to the screen for changes; applicable data entitlement and Normal usage
- **Audience:** Portrait users operating saved screens
- **Surfaces:** web; ios
- **Availability:** Production with account and usage requirements
- **Content type:** how-to
- **Last reviewed:** 2026-09-22

Use the Screening list to manage a saved screen and open its execution history. You need edit access to change filters, run the screen, or change its schedule. Read access lets you inspect shared work; use the shared-copy action when you need your own editable configuration.

Create, edit, and run screens on the web and iOS. Use the web app for automatic-run schedules, sharing, duplication, cancellation, and deletion. The management instructions below refer to the web app; mobile execution and export are described separately.

## Run and verify

1. Open **Screening**, then open the screen or choose its **Run screening** action. For a new screen, review the builder's filters before choosing **Run screen**.
2. Confirm the universe, expressions, qualitative windows, and notification choice. See [Filter reference](https://portraitresearch.com/documentation/screening/filter-reference/) for units and missing data.
3. Start the run. Normal usage and applicable concurrency checks can refuse it before execution. In the new-screen flow, the configuration is saved before the run request, so a refused run can leave a saved screen with no new run.
4. Expand the screen's history and open the intended run. Check its timestamp and filter snapshot, then inspect the funnel and company-level evidence. A completed run with no matches is different from a failed or cancelled run.

Use **Rename run** in history when a name helps distinguish experiments. Renaming does not recalculate results. To change conditions, edit the saved screen and start a new run; preserve the earlier run when you need to compare scopes.

| State | What to verify or do |
|---|---|
| Queued / pending | A run exists but has not started processing; wait or request cancellation |
| Running | Inspect progress; results may be incomplete |
| Completed | Review the final filter funnel, matches, missing data, and evidence |
| Failed | Inspect the displayed failure and scope before trying again; a later attempt is a separate run |
| Cancelled | Treat retained output as partial; verify that cancellation has finished before deleting the run |

### Run and review on iOS

Open **Screening → New screen**, or open a saved screen and choose **Filters → Edit filters**. Build or refine the filters in the conversation, confirm additions and updates, and tap **Run screen**. See [Build a screen with the AI Assistant](https://portraitresearch.com/documentation/screening/build-with-ai-assistant/#build-or-edit-a-screen-on-ios).

The result page shows run history and company results. Open a company to inspect its filter results and available evidence. Use **Export** to download the current result set as CSV, or **Create project** to continue research with those companies. Mobile uses the full current result set for these actions; selecting a subset is a web control.

## Dates depend on how you start the run

| Action | Date behaviour |
|---|---|
| **Run screen** from the new-screen or edit-screen builder, on web or iOS | Uses the saved filter dates without advancing them to the run date |
| Run again from the web list or result page | Refreshes qualitative windows and supported rolling quantitative return windows before recording the new run |
| Automatic daily or weekly run | Applies the same date refresh before recording the scheduled run |

Qualitative windows preserve their saved length in days and end on the new run date, including windows entered as fixed dates. Quantitative dated expressions advance only under the supported rolling rules. Review [Qualitative evidence dates](https://portraitresearch.com/documentation/screening/portrait-textual-reasoning/) and [Quantitative recurring screens](https://portraitresearch.com/documentation/screening/standard-quantitative-filters/) before relying on a historical comparison.

## Set automatic runs

In the screen's **Automatic runs** control, choose **Off**, **Daily**, or **Weekly**. For weekly runs, select the day as well as the hour. Confirm the displayed next run after saving.

The picker displays the viewer's local time and converts it to a stored UTC day and hour. It does not store a named time zone with daylight-saving rules. Review the displayed schedule after a daylight-saving change or when working in another time zone if keeping the same local hour matters.

Scheduled runs use the screen creator's access and usage policy. An inactive creator is skipped. If the budget check refuses a scheduled run, that attempt creates no run and queues no work; do not interpret a due schedule as evidence of completed coverage. Scheduled admission does not use the same interactive concurrency check as a manual run, and it does not allow budget overage.

The completion-email option can notify everyone with access to the screen. Review it before enabling a recurring run. **Off** stops future automatic runs; it does not cancel a run already in progress.

## Copy, cancel, and delete

- **Duplicate** creates another saved screen. Duplicating your own screen can retain its schedule; copying a screen owned by someone else starts without that schedule. Review **Automatic runs** on the copy before leaving it unattended.
- Request **Cancel** for an active run, then verify its terminal state. Cancellation is a request, not an immediate erasure of work already produced.
- **Delete run** removes that execution from history and makes its results unavailable through the run. There is no restore action for the deleted run. Cancel an active run before deleting it. Other runs and the saved screen remain available.
- The screen's **Delete** action removes the saved screen from the active list. Use **Off** when your intention is only to stop future automatic runs.

## Share a screen

Use **Share** on your screen's row to give selected teammates read-only access. Follow [Share with teammates](https://portraitresearch.com/documentation/documents/sharing/#share-with-teammates) for permissions and recipient management.

Recipients can [find shared work and create an independent copy](https://portraitresearch.com/documentation/documents/sharing/#find-work-shared-with-you). A copied screen does not inherit the original's automatic-run schedule.

## Export or continue the research

On the web result page, **Export to Excel** downloads a `.csv` file with company name, ticker, and the available reasoning for each filter. Select companies to export that selection; with none selected, the export uses the current result set. On iOS, **Export** downloads the current result set in the same CSV format, which opens in Excel.

On the web, select matching companies to create a Project for follow-up research. On iOS, **Create project** uses the current result set. Check the company set and name before saving the Project.

## If a run is missing or incomplete

Check run history first. If no new run exists, inspect the visible error, **Settings → Usage**, screen permissions, and data access. A missing scheduled run can also reflect an inactive creator or refused budget admission. If the run exists but failed or stopped, inspect its status before starting another attempt.

For an unresolved problem, [contact support](https://portraitresearch.com/documentation/more/contact-us/) with the screen link, run timestamp, time zone, and safely redacted error. Do not include credentials or licensed source text.

## Screening and Scheduled Research

A scheduled screen reruns its saved filters. [Scheduled Research](https://portraitresearch.com/documentation/scheduled-research/overview/) instead runs a saved Easel investigation and has separate authorization, pause, usage, and notification rules. Identify which workflow produced the run before diagnosing a missing email.