# Watch for topic spikes and escalate > Scheduled fresh-upload sweeps detect when your keyword's video volume jumps; chat drafts the alert with context. - For: Analyst - Category: Agent & Workflow Automation - MCP tools: deep_search, chat - HTML version: https://veridive.com/mcp/use-cases/topic-spike-watcher ## The problem Stories have a video signature: before a topic dominates the cycle, its upload volume moves — explainers appear, commentary channels pick it up, the count of fresh videos mentioning the keyword jumps from two a day to twenty. Whoever watches that gradient hears the story building while it's still a whisper in text media. Nobody watches gradients manually; watching is a machine's job. A scheduled sweep that counts and characterizes fresh uploads on your watch terms — and escalates only when the pattern breaks — is an early-warning system that costs prompts and pays in lead time, whether the topic is your company, your category or your risk register. ## How veridive MCP helps **deep_search** with the `today`/`this_week` upload filters is the sensor: each scheduled sweep returns what's new on the watch terms — volume, sources, and what the transcripts actually claim. Comparing sweeps is the detector: today's count and tone against the baseline, drift versus spike. **chat** is the escalation layer: when the pattern breaks, it drafts the alert *with context* — what jumped, who's driving it, what the new videos claim, why it might matter — instead of a naked "volume up" that still requires the human to do the reading. Quiet days cost a sweep; loud days buy you the hours between the whisper and the cycle. ## How to run it 1. **Connect veridive MCP** — One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes. 2. **Define the watch and the baseline** — The terms that matter and a first sweep to calibrate normal. ``` Sweep videos uploaded today mentioning — count, sources, and what each claims. This is our baseline; remember its shape. ``` 3. **Schedule the sweep-and-compare** — Daily (or hourly, for hot topics): new sweep, compared to the baseline, escalate only on breaks. ``` Schedule daily: repeat the sweep and compare to previous days. If volume or tone shifts notably, draft me an alert; otherwise reply 'quiet'. ``` 4. **Get alerts with the reading done** — The escalation arrives as a briefing, not a metric — what, who, claims, why it matters. ``` For today's spike: summarize what changed, who's driving the coverage, and the main claims — with links to the key videos. ``` ## Problems it solves - Topic build-up is visible in upload gradients that no human watches manually. - Text-media monitoring hears stories after the video layer already moved. - Naked volume alerts still require the human to do all the reading. - Continuous vigilance on a risk register doesn't fit anyone's calendar. ## Why it's valuable - Lead time: the hours between the whisper and the cycle, bought with a scheduled prompt. - Escalations arrive pre-read — what, who and claims, not just a number. - Quiet days are cheap; the watcher scales to a whole risk register of terms. - Flips into crisis-monitoring mode seamlessly when a spike is about you. ## Example prompts ``` Baseline sweep on : today's uploads, count, sources, claims. ``` ``` Schedule daily sweep-and-compare; alert on breaks, else 'quiet'. ``` ``` For the spike: what changed, who's driving it, main claims — with links. ``` ## FAQ **How does the agent 'remember' the baseline?** Via the conversation or your client's memory for scheduled tasks — each run's summary becomes the next run's comparison input. The prompt carries the state; no external infrastructure needed. **What counts as a spike worth escalating?** You define it in the instruction — notable volume jumps, new prominent sources, or claim shifts. Start loose, tighten after a week of false positives; the tuning is one sentence. **How does this relate to crisis monitoring?** This is the standing sensor; crisis monitoring is the active mode after a spike about you fires. The watcher's alert is often the thing that starts the crisis workflow with those crucial extra hours. ## Related use cases - [Monitor a crisis as it unfolds on video](https://veridive.com/mcp/use-cases/crisis-video-monitoring.md) — Fresh-upload sweeps plus comment sentiment show how the story and the reaction are moving hour by hour. - [Schedule a morning intelligence briefing](https://veridive.com/mcp/use-cases/morning-intel-briefing-agent.md) — A scheduled agent checks the connection, sweeps the last day's uploads in your space, and drops summaries before your first coffee. - [Follow regulation that hits your industry](https://veridive.com/mcp/use-cases/regulation-radar.md) — Filter to fresh uploads from legal and policy channels, then get a grounded brief on what's changing and who says it matters. --- Part of the veridive MCP use-case library: https://veridive.com/mcp/use-cases.md Server docs for agents: https://veridive.com/mcp.md