# Will Francis: Proof-led walkthrough A result-first walkthrough with numbered steps, highlighted prompts and clear before-and-after proof. Harper’s editing structure, interpreted for Will. Version 1.0.0. Reviewed 2026-10-15. ## Use for Workflow tutorials, Before & after demos, Practical case studies. ## Structure & pacing Give the viewer a result before the method. Observed: The reference opens with thumbnail evidence, moves through numbered steps and returns to the resulting workflow. Wider presenter shots separate denser screen-led sections. - Open with a real before-and-after or a tangible result relevant to Will’s subject. Name the specific finding. - Follow the actual process with as many steps as it needs. The source’s five steps are not a universal template. - Alternate a comfortable presenter shot with proof-led compositions. Let each step have one visible action and result. - Finish on the achieved outcome and one useful next action, retaining any caveat. Starting settings (chosen production guidance, not extracted values): Suggested rhythm: a visual update around a new action or phrase, often 2–4 seconds. Dense prompts and results need longer holds. Fit the script, not a fixed duration. ## Colour & footage Bring the editorial warmth into Will’s studio. Observed: Harper’s source is a warm domestic setting paired with light graphic cards. That location and palette belong to the source creator. - Keep Will’s own studio, wardrobe and natural skin tones. Do not recreate Harper’s room or identity. - Use warm off-white #F6F4EE for evidence panels, ink for text and yellow for the currently important detail. - Preserve neutral screen whites and the source footage’s real background lighting. Warmth comes primarily from the designed panels. - Match camera exposure across crop changes. Avoid a blanket orange or teal grade. Starting settings (chosen production guidance, not extracted values): No source LUT is claimed. Start from a neutral camera correction; compare face, paper whites and studio lights across scenes. ## Captions Separate what is said from what is shown. Observed: The reference uses sentence-case sans subtitles and separate editorial graphic headings. Prompt highlights guide attention within a larger card. - Keep Will’s Avenir Next Bold subtitles, using short phrases and a charcoal backing for legibility. - Use Sentient for graphic headlines, Inter for UI labels, and sentence case throughout. - Position subtitles consistently over an uncluttered presenter area. Keep them away from the highlighted prompt. - Use yellow emphasis within the prompt card, not karaoke-style highlights on every subtitle word. Starting settings (chosen production guidance, not extracted values): At 1080 × 1920, start captions at 56–64 px and keep two lines maximum. A step badge is functional navigation, not a decorative label above every heading. ## Graphics & evidence Make the process visible, one action at a time. Observed: The source combines before/after thumbnail grids, numbered step cards, a prompt with moving highlights, screen recordings and a results dashboard. - Borrow the result-first order, numbered steps and selective prompt highlighting. Build new layouts from Will’s palette and type. - Use real screen recordings in large bordered panels. Move the presenter below the evidence when the demonstrated action needs the upper frame. - Highlight one prompt clause only as it is explained. Show the resulting output immediately afterwards. - Only include counters or comparisons supported by real input/output. Never reuse the source’s 34 videos, five thumbnails or percentages as Will’s results. Starting settings (chosen production guidance, not extracted values): Start with an evidence panel occupying about 55% of the vertical frame. Use a small yellow step chip, an ink border and an uncluttered cream ground. Abstract diagrams still use the current explainer skill. ## Sound & transitions Keep the energy in the explanation. Observed: Clear spoken dialogue drives the reference. The viewing analysis does not establish a specific music track or repeatable sound-effects recipe. - Keep Will’s original narration as the primary track and cut out repeated or empty phrases. - Use straight cuts for a new step, with brief restrained movement inside a panel to show an action. - Default to silence under the voice. Add licensed, quiet music only if the particular video benefits. - Do not imitate a creator’s voice, copy a soundtrack or insert a sound effect on every step. Starting settings (chosen production guidance, not extracted values): Suggested delivery target: around −16 LUFS integrated with true peak at or below −1 dBTP. Listen through on headphones and phone speakers. ## Sources and evidence limits Derived from the supplied 73-second Harper Carroll reference. Its editing mechanisms are adapted to Will’s identity. Font names, exact colours, grade and audio settings cannot be recovered reliably from a viewing analysis. - [Harper Carroll · ChatGPT thumbnail workflow](https://www.instagram.com/p/Deev3b0s29A/). Supplied reference. 73-second reference. Scene-by-scene visual analysis completed 15 October 2026. Creator assets are linked, not included in the style pack. 00:00–00:04 before proof · 00:09–00:24 numbered steps and prompt · 00:40–00:47 dashboard · 01:01 testing proof ## Example brief Use the Will Francis Proof-led walkthrough style, adapted from the linked Harper Carroll reference. Read the attached recipe and current brand guide. Open with a real result, explain the workflow in numbered steps, highlight one useful prompt clause at a time and finish with visible proof. Use my footage, Sentient headings, Inter UI text, Avenir Next Bold captions, cream evidence panels and yellow emphasis. Keep natural skin tones and my studio colour. Use only real results and source recordings. The reference supplies editing structure, not its creator’s identity, music or claims. Plan source timecodes, caption breaks and missing evidence before rendering. # Shared brand rules Read /Users/willfrancis/Library/CloudStorage/Dropbox/Claude/brand-will-francis/BRAND.md before asset creation. - Keep Will’s voice, natural skin tones and recognisable studio. - Use Sentient for headings, Inter for interface text and Avenir Next Bold for subtitles. - Use authentic screenshots and source footage. Never fabricate a product result or performance metric. - For narrated HTML diagrams, use the current explainer-slideshow skill, its 800 × 800 template and verification scripts. - Keep source creator footage, music and identity out of the downloadable pack. If the folder is unavailable, ask for the brand materials. The current narrated HTML explainer skill at /Users/willfrancis/.codex/skills/explainer-slideshow/SKILL.md overrides older slide rules. Separate raster covers use GPT image generation with supplied brand references.