Hire an AI YouTube title and retention reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A dozen or more title and thumbnail pairs, each changing one thing
• The first minute of the video, written to pay off the promise
• Three titles that differ enough to test, with a way to test each
• The place where viewers leave most, what you said then, and one change

How it does the job well

• Write for one named audience and one action. Each asset has a single job.
• Use the customer's own words from reviews, calls or surveys before inventing language.
• Every claim needs proof the business can show. Leave unproven claims out or mark them.
• Judge results against the business goal, such as profit or qualified leads, not clicks alone.

Mistakes it avoids

• Generic copy that could fit any brand.
• Comparing periods or campaigns that are not comparable, such as different dates, budgets or tracking.
• Calling a test winner before there is enough data.

What a good result looks like

An example, not your task: You have a video idea and want the title and thumbnail to earn the click.

• Step 1: Pairs. The single promise, 12 or more title and thumbnail pairs that each change one thing, and a plan for the first minute.
• Step 2: Three to test. Scores on five rules, the titles cut and why, the three picks and how to test each.
• Pause to publish and export. You publish the top pair, wait, then download the retention report from YouTube Studio.
• Step 3: Transcript. The video's words with times, from the live video or from captions you paste.
• Step 4: Where viewers leave. The loss in the opening, the steep drops matched to what you said, the slow slide and the one change to make.
• Hand-back summary. Each step's file and every MISSING line in one list.

How the work is checked

• The title and thumbnail make the same single promise.
• Each pair differs from the base by one change, so a result can be read.
• The three picks differ in type or reason, not in wording alone.
• Each drop in the retention report is matched to words from the transcript.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way MrBeast (Jimmy Donaldson) and John Caples would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Plan the pairs. Writes the single promise, then at least twelve title and thumbnail pairs that each change one thing, and the first minute that pays off the promise.
• Cut to three. Reads each title with the video hidden, scores it on five rules, drops weak ones and keeps three that differ from each other.
• Publish and export (you). You publish the top pair. After a week or more you download the audience retention report for the video from YouTube Studio.
• Get the words of the video. Paste the video's captions from YouTube Studio. This is free and the default. The transcript reader can fetch them for you instead, but it uses a paid service (TranscriptAPI) that needs your own key and has a small free allowance.
• Read where viewers leave. Finds the loss in the first 30 seconds, the steep drops and the slow slide, matches each to what you said, and names the one biggest leak.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id pick-a-title-and-thumbnail-then-fix-where-viewers-leave only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/pick-a-title-and-thumbnail-then-fix-where-viewers-leave/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Your video idea and what the viewer gets
• Your channel, style and audience
• Past results and lessons (optional)
• The link to the video once published
• The audience retention export from YouTube Studio

Tools and access

• Any good chat AI: Writing pairs and a first minute is text work any strong chat AI does well with this skill.
• YouTube Studio: Its click rate and audience retention reports show which pair won and where viewers left.
• Any good chat AI: Reading, scoring and ranking lines is text work a strong chat AI does well with this skill.

Functions and deliverables

• YouTube title and retention reviewer: Share a video idea, then real viewing data and captions. Get title and thumbnail options, plus a review of where viewers leave.
  • A dozen or more title and thumbnail pairs, each changing one thing
  • The first minute of the video, written to pay off the promise
  • Three titles that differ enough to test, with a way to test each
  • The place where viewers leave most, what you said then, and one change
• Video title and thumbnail planner: Plan your video title and thumbnail pairs before the shoot. Get a dozen test pairs that change one thing each, plus a first minute that pays off.
  • A dozen or more title and thumbnail pairs, each changed by one thing
  • A one-idea description of each thumbnail
  • The first minute written to pay off the promise
  • A result log, ready before the first test
• Headline and cover test planner: Share headline or cover options and verified offer facts. Get a ranked shortlist, clear reasons and a response test to compare your strongest options.
  • For each line, what a stranger would think it offers
  • A score on five rules, with a reason for each
  • The lines cut outright and the rule each broke
  • Three picks that differ, and how to run an outside read
• YouTube transcript and channel collector: Give it a video or channel. Get transcripts, video lists and new uploads. Collects data from websites.
  • A timestamped transcript
  • A list of matching videos
  • The latest uploads with views and dates
  • Search hits inside a channel
  Source: Alireza Rezvani.
• YouTube viewer drop-off reviewer: Turn a YouTube retention CSV into the biggest viewer drop and one suggested fix.
  • A hook leak reading
  • Cliff points with what you said there
  • A note on any steady slide
  • The one change to make first
  Source: Jake Schincariol.
• Brand profile builder: Share your website. Get a reusable profile covering your products, customers, voice and approved claims, with sources.
  • A folder of eight plain files, from a start file for your AI to a claims list and a change log
  • A tag on every fact, and every claim marked approved, pending or banned
  • A two-sentence diagnosis of your biggest gap, and a list of what's missing and who can answer it
  • A short slice for your next step, if you name one
• MrBeast (Jimmy Donaldson): Plan a video idea using MrBeast’s recorded approach. Match the title and thumbnail to the content, review the opening and check critical production needs.
  • Your problem worked on the way MrBeast would: their questions, their method, their push-back
  • The answer or the work, in MrBeast's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step
• John Caples: Write headlines and plan a fair ad test through John Caples’s recorded approach. Lead with the reader’s gain and count replies by version.
  • Your problem worked on the way John Caples would: their questions, their method, their push-back
  • The answer or the work, in John Caples's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• You have a video idea and want the title and thumbnail to earn the click.
• Your videos get views but viewers leave early and you don't know why.
• You want to pick the next video from what worked, not from a hunch.

Limits to check before starting

• You only want title and thumbnail ideas and have nothing published yet: use the Video title and thumbnail planner.
• You want the first seconds of a video ad scored before you film: use the Video hook scorer.
• Your video has too few views to show a pattern: wait for more views before reading the retention report.
• You want views from ads on YouTube read, not a video's own viewers: use the YouTube ad results reader.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Download the handoff pack. It's a zip with HANDOFF.md and every skill file this workflow uses.
• Give the folder to Claude (in Claude Code, or as files in a Claude Project) and say "Follow HANDOFF.md".
• Describe the video and your channel. Choose one of the three pairs and publish it.
• After a week or more, open YouTube Studio, download the audience retention report for that video and give it back with the video link.
• Paste the video's captions from YouTube Studio too. This is free. The workflow can fetch them with a paid service (TranscriptAPI, your own key) only if you say yes.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/pick-a-title-and-thumbnail-then-fix-where-viewers-leave for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.