Keep an Elicit research watch running after 31 October

Existing Elicit Alerts stop on 31 October 2026. Their replacement, Routines, requires Pro, Scale or Enterprise. Old Alerts did not consume monthly usage; each Routine run does.[1][2][3]

This guide follows documentation checked on 29 October. The example brief below is an original exercise; it has not been run in Elicit.

Preserve and convert

1. Export any old results you want as CSV before 31 October, using the download button above the Alert’s results.
2. Open the Alert and choose Convert to Routine. Review the proposed instructions and schedule; conversion switches off that Alert.
3. Convert each Alert separately. Nothing migrates automatically.
4. Choose a daily or weekly schedule in your local time, and describe the output you want.

Reviewing the proposal uses no allowance; the first run does. Task size and complexity affect usage. Check Account Settings for your balance and Run history for results. Pause a Routine to stop future runs.[1][2][3]

Give the watch a precise brief

Suppose you are comparing passive cooling in occupied school classrooms. A broad search for “cooler schools” could collect material that answers quite different questions. Adapt this proposed brief to the decision you face:

Find new research on passive cooling in existing, occupied school classrooms. Include measured building studies; separate simulation-only papers. For each eligible study, record its DOI or stable link, version, publication date, location, intervention, comparator and measured outcome. Distinguish newly published papers from older papers newly found. Explain one finding that could change our comparison, with its limitations. Flag corrections or retractions. If nothing qualifies, say so.

Read one report before committing

Inspect the papers behind the first report. Does “cooler” mean lower classroom temperature, fewer hot hours or a reported feeling of comfort? Keep those outcomes separate. Check whether an intervention was measured in an occupied room, and whether the comparator faced similar weather. A citation should let you examine the actual comparison.

Keep a short record of what the report found and what you changed in the brief. An empty report is useful only if the search was adequate; it cannot establish that no relevant study exists. Continue the watch when its updates improve a real decision.

Report an error or suggest a correction