Stop asking for five generic bullets. Compare the video with your real context and decide what deserves action.
Analyze this YouTube video against my real operating context. The goal is not to summarize the video. The goal is to decide what matters to me, what does not, and what deserves action. Video: [YOUTUBE URL] Use these context sources if they are available: - who I am and how I work; - company and product map; - current priorities and active projects; - software stack and existing processes; - constraints, non-negotiables, and approval boundaries; and - recent decisions that affect this topic. If you cannot access the video's supported transcript, ask me to paste it. Treat the transcript as text evidence only. Flag any point that depends on a visual demonstration you cannot see. If a required context source is missing, list it. Do not guess how my company works. Produce five sections: 1. Source map - the video's main claims, examples, and recommendations; - transcript support for each important claim; and - anything uncertain, promotional, or dependent on unseen visuals. 2. Relevance map For every material idea, show: - why it matters or does not matter to me; - the company, project, process, or tool it touches; - what I already do in this area; - whether the idea confirms, challenges, or extends that practice; and - one label: use now, test, investigate, remember, or ignore. 3. Change queue Separate: - changes worth making now; - small tests that should come first; - questions that need better evidence; and - interesting ideas that are irrelevant to my current priorities. 4. Consequence check For every proposed change, name the expected benefit, cost, affected people or systems, main risk, and the evidence that would show whether it worked. 5. Next paths Offer up to four bounded next steps: research it, make a plan, build a small test, or implement it. State exactly what each path would do and what approval or information it needs. Stop after presenting the paths. Do not change files, tools, schedules, or external systems until I choose one.
Replace anything in [brackets] with your real thing. The prompt is deliberately short — resist the urge to pad it.
When a YouTube video may contain something useful, but a generic summary cannot tell you what should change in your actual work.
A transcript summary tells you what the creator said. This compares the evidence with your priorities, tools, decisions, and constraints—then separates real change from interesting noise and stops before acting for you.
Transcript access is not visual understanding. Flag claims that depend on a demonstration the AI cannot see.
The approved walkthrough is intentionally illustrative: a transcript passes through Matthew’s company brain, separates what is new, different, worth testing, already better, or safe to ignore, and ends with bounded next paths.
The video for this one drops on @espoai — the page is live so the prompt is yours already.
The same method, pointed at the situations people actually hit.
Analyze this YouTube video against my real working context: my role, current goals, active projects, tools, constraints, and recent decisions. Do not give me a generic summary. Show what is new, what challenges my current approach, what deserves a test, and what I should deliberately ignore. Stop before taking action. Video: [URL]
Analyze this YouTube video against our team context: [CONTEXT FILES OR NOTES]. Produce a source map, relevance map by team/process, a small-test queue, risks and affected owners, and up to four next paths. Do not assign work or change systems until an owner approves a path. Video: [URL]
Process this saved YouTube video against my current priorities and existing methods. Tell me what is genuinely new, what we already do, what needs better evidence, what belongs in a later reference list, and what can be discarded. Video: [URL]. Context: [FILES OR NOTES].
Every method gets a video on @espoai and a page here. Subscribe for the next one, or get the weekly email with what I’m actually using.