A Reel with a big number isn’t automatically a useful trend for us. We need a topic, a time window and a reason to study the example. We’ll use pottery process footage to make that distinction concrete: a glossy product reveal and a close-up teaching moment might need completely different assets.
Kabo, an AI agent for creator research, finds Reels and available account context. Muse helps us accept or reject candidates before we start a separate creative brief.
Connect Kabo before asking for fresh evidence
Use the Muse connection guide to connect https://kabo.sh/mcp-pub and sign in. Ask Muse to show the ready Kabo tools before requesting fresh research.
Prepare the topic and shortlist rules
[ ] Topic — The specific making step or audience problem, not just the niche. [ ] Time window — Exact start/end publication dates and retrieval date. [ ] Owned footage — Process, reveal, talking head or customer use; say what is missing. [ ] Candidate rules — Relevant language and format; exclude reposts or duplicates when identifiable. [ ] Decision — A short list to inspect, not a promise about future reach.
Agree on what belongs in this shortlist
Our potter needs references about a visible making step, not customer reveals. In the first message, fix the topic, exact publication window and format. Ask Muse to show candidates before discussing which one to study. A beautiful example outside the dates can be a separate style reference, but it should not silently join the recent sample.
Use Kabo to find recent public Instagram Reels about [specific topic], published between [start date] and [end date]. Show up to 6 candidate URLs and why each is relevant. I have [owned footage] but not [missing footage]. Don’t rank candidates by raw views alone.
Ask for an evidence card for each accepted candidate
After removing irrelevant candidates, have Muse use Kabo to add the available metrics and recent account context. Ask for two separate judgments: is this relevant to our making question, and is there enough account context to call it unusual for that creator?
A missing baseline changes the second answer without automatically erasing the first.
For the accepted Reel URLs, use Kabo to add available public metrics and recent same-account context. Return a card for each: URL, publication date, retrieval date, relevance reason, baseline coverage and status: supported relative breakout, interesting reference with unresolved baseline, or exclude. Explain the status using returned evidence. Do not develop original concepts yet.
Use the next message to remove an unhelpful mechanism
If the shortlist is dominated by finished-product reveals, tell Muse that your available footage shows joining and finishing. Ask it to keep only references that help answer a process question, or explain why no useful recent example was found. The right outcome can be a narrower sample.
Stop when we have references worth inspecting
The deliverable here is a reviewed shortlist with reasons to keep, exclude or defer each candidate. Ask Muse which uncertainty prevents a confident comparison, then decide whether another lookup is worthwhile.
When we have chosen a reference, the Muse ideas guide takes over. Its job is to build our original viewer decision and demonstration; this shortlist should not grow a second filming plan.
A reference can be useful without winning a ranking
An illustrative decision card might read:
- Relevance: close-up handle attachment addresses the process we want to study.
- Account context: unavailable in this run.
- Decision: keep as an explanation reference; do not label it an account-relative breakout.
- Next check: inspect whether the visible making step actually answers our question.
We have not assigned a fake URL or view count. Fill those fields from the actual lookup before using the card.
This Reel has no usable recent account baseline in the returned evidence. Keep its relevance note, remove the relative-performance claim and show the revised card. Which remaining comparison, if any, is still supported?
Open one reference before planning the shoot
Copy the accepted cards into our research notes, with the dates and unresolved checks intact. Use the Instagram comparison guide for a broader peer study, or move to original ideas once the reference choice is made.
Worked examples are illustrative. Product screenshots show existing Kabo web results, not a run of the example in this guide. Use your own materials and the evidence returned in your task.