Lesson 05 · Activity 03

Scheduled Integrity Monitor

Research goes stale. A figure that was current when you cited it can quietly become wrong. In this activity you'll build a recurring Task that re-runs your query on a schedule and delivers a fresh, cited summary — then you'll practice the synthesis step itself, triangulating what multiple sources say in a structured matrix instead of trusting any single one.

Part A — Build the recurring Task

Fill in the fields and generate a Task prompt. In Perplexity, create a scheduled Task (a Pro feature) and paste it in.

Topic to monitor

How often should it run?

Only include items from the last…

What should each summary prioritize?

Fill in the fields above, then choose "Generate Task prompt" to build your scheduled-Task instruction here.

Tasks turn Perplexity from reactive to proactive: it runs on schedule and notifies you when a fresh, cited briefing is ready — so monitoring becomes a maintained record, not a one-time snapshot.

Part B — Synthesis matrix

When your Task (or a Deep Research run) returns several sources on one question, don't just take the top answer. Lay the claims side by side. Add a row per claim, note what each source says, and mark whether they agree. The matrix is saved on this device.

Claim / question	Source A says	Source B says	Verdict	Remove row

Synthesis write-up

In 3–5 sentences, what can you say with confidence, and where do the sources genuinely conflict?

Saved automatically on this device.

The integrity payoff: Agreement across independent sources earns confidence; conflict is a signal to dig, not to pick the convenient answer. A scheduled Task keeps the record current, and the matrix forces the triangulation that turns "the AI said so" into "here's what the evidence shows."