Jeff Bullas

Nice focus — wanting a clear, beginner-friendly workflow is exactly the right place to start. Below is a practical, do-first guide that gets you a visible result fast and walks you through creating true embroidery files.

Quick win (under 5 minutes): Open a simple AI-generated PNG in Inkscape, use Trace Bitmap to turn it into an SVG, simplify it to bold shapes, and save the SVG. That SVG is your foundation for digitizing into an embroidery file.

What you’ll need

• AI image (PNG with transparent background works best)
• Inkscape (free) to vectorize and simplify
• Ink/Stitch (free plugin for Inkscape) or a beginner embroidery program (Embrilliance/SewArt/Wilcom)
• A machine format you need: DST is universal; PES is common for Brother machines

Step-by-step workflow

1. Open your AI PNG in Inkscape.
2. Use Path → Trace Bitmap. Choose Brightness or Edge detection and preview. Click OK to create a clean vector.
3. Ungroup and delete small specks. Simplify nodes (Path → Simplify) until shapes are bold.
4. Limit colors: reduce to 2–4 flat colors. Convert complex gradients to flat fills.
5. If using Ink/Stitch: select shapes and set Stitch Type (Satin for outlines, Fill for areas). Use Ink/Stitch Parameters to set stitch density and underlay.
6. Run Ink/Stitch → Embroider or Export to DST/PES. If using another program, import the SVG and use its digitize tools to assign stitch types and export a machine file.

Example

Generate a simple hummingbird: vectorize, remove tiny details, make the beak a satin stitch, wings as fill stitches, export DST. Put on a test scrap fabric at low speed to check stitch density.

Common mistakes & fixes

• Too many colors or gradients —> simplify to flat colors.
• Thin lines that disappear —> thicken outlines to 1.5–2 mm equivalent.
• Excessive detail —> remove small elements or merge them into larger shapes.
• Incorrect stitch density —> lower density for thicker fabrics, increase for delicate fabrics; test stitch.

Copy-paste AI prompt (use this to generate embroidery-friendly art):

Create a simple, high-contrast image suitable for embroidery: flat vector style, no gradients, bold outlines, maximum 3 colors, large simple shapes, transparent background, 3000×3000 PNG. Subject: [your subject here].

Action plan (next 30–60 minutes)

1. Generate or pick an AI image and open it in Inkscape.
2. Vectorize, simplify, limit colors, and export SVG.
3. Install Ink/Stitch or open your embroidery software and import SVG.
4. Digitize (assign stitch types), export DST/PES, and test on scrap fabric.

Keep it simple at first. Start with a single-color logo or icon and build confidence. Embroidery is forgiving if you design for the stitch — design fewer, bolder shapes and test early.

Nice focus — the real win is treating translation and synthesis as two separate, repeatable steps. That makes the work faster and more reliable.

Here’s a practical, step-by-step way to translate and synthesize non-English research so you can act on the findings quickly and confidently.

What you’ll need

• PDF or scanned paper (digital copy). If scanned: OCR tool (many PDF readers do this).
• An AI assistant with strong language skills (e.g., GPT-4‑level) or a high-quality translator (DeepL) for fidelity.
• Note-taking tool or reference manager (Zotero, Mendeley) to store citations.
• Time: 30–90 minutes per paper for high-quality work.

Step-by-step process

1. Extract the text: run OCR if needed, or copy the digital text into a document.
2. Quick skim in original language: note headings, figures, and any unfamiliar terms.
3. Translate sections, not the whole paper at once: start with title, abstract, conclusions, then methods & results.
4. Ask the AI to produce three outputs for each section: a literal translation, a plain-English paraphrase, and a 3‑bullet takeaway.
5. Synthesize: combine section takeaways into a structured summary (background, key findings, methods, limitations, implications).
6. Verify: ask for bilingual checks on critical sentences—compare original vs translation for nuance.
7. Save: store the original, translation, and synthesized notes with citation metadata.

Copy-paste prompt (use as a base)

Translate the following [insert section: abstract/introduction/results] from [Original language] to English. Provide three outputs: (1) a literal translation, (2) a plain-English paraphrase that a non-specialist can understand, and (3) three concise takeaways with confidence level (high/medium/low). Also flag any technical terms or ambiguous phrases.

Prompt variants

• Brief summary for executives: “Summarize this paper in 5 bullet points with one sentence on why it matters.”
• Technical verification: “Compare the original sentence and translation. Highlight any possible mistranslation and suggest alternatives.”
• Lay explanation: “Explain the main results as if to a smart 16‑year‑old.”

Common mistakes & fixes

• Literal, awkward translations — fix: request both literal and paraphrase versions.
• AI hallucinations about data or methods — fix: ask to quote original sentences and mark where the model is uncertain.
• Missing figures/tables — fix: extract captions and table text separately and translate them.

7‑day action plan (quick win)

1. Day 1: Pick one important non-English paper.
2. Day 2: Extract text and run translations for abstract + conclusions.
3. Day 3: Produce a 1‑page synthesized summary and store it.
4. Day 4: Verify key terms with a bilingual check and refine.
5. Day 5–7: Repeat for two more papers—build a small translated library.

Start small, build a template, and you’ll speed up every paper after the first. Want a tailored prompt for a specific language and field? Tell me the language and topic and I’ll draft it for you.

Best, Jeff

A common concern, though the term ‘shadowban’ itself is often misunderstood.

Short Answer: While X does not have an official ‘shadowban’ feature, you can check for reach limitations by logging out and searching for your account and posts; however, reduced visibility is almost always tied to your content strategy.

Let’s look at the specific content format habits that can cause the platform to algorithmically limit your reach.

Sustained low reach is less about a hidden penalty and more about the quality signals your content formats are sending. First, if you consistently post low-quality video or blurry image formats that fail to earn engagement, the algorithm learns to deprioritise your content over time because users are ignoring it. Second, using a repetitive text-based format, such as posting the same link with identical copy over and over, is a classic spam signal that will severely restrict your visibility. Finally, any content format that is frequently reported by other users, regardless of whether it technically violates a rule, can be subject to algorithmic suppression. Instead of worrying about a hidden ‘ban’, you should focus on creating high-value, engaging content formats, as this is the only reliable and ethical way to ensure your posts are seen.

Cheers,

Jeff

Thanks — that’s a practical, useful question about starting with quick wins when building SOPs for new tools.

Here’s a simple, confident path you can follow today. The idea: start small, use AI to draft the heavy lifting, then test and improve with real people.

What you’ll need

• Clear goal: what the SOP must achieve (eg. get a new hire productive in 2 days).
• One or two subject-matter people to review steps.
• Access to an AI writer (ChatGPT or similar) and a document editor (Google Docs, Word).
• Example screenshots or short video of the tool (optional but helpful).

Step-by-step — build your first SOP

1. Define scope: Pick one tool and one role. Keep scope tiny (eg. onboarding the marketing team to Tool X).
2. Map the current steps: Ask one person to show you how they do it. Note 6–12 steps, pain points, and time taken.
3. Ask AI to draft: Use the prompt below to generate a clear SOP draft with checklist, estimated time, and troubleshooting tips.
4. Review with an SME: Send the AI draft to the person who does the work. Ask for corrections and missing steps.
5. Test with a new user: Have a new person follow the SOP and time them. Collect two small changes.
6. Lock and publish: Finalize the document, add screenshots, and store it where your team finds it.
7. Iterate: Re-run the process after 30 days or after a version change in the tool.

AI prompt (copy-paste)

Use this prompt in your AI tool. Replace bracketed text with specifics.

“Create a step-by-step SOP for onboarding [ROLE] to use [TOOL NAME]. Include: purpose, prerequisites, estimated time, detailed steps (with screenshots placeholders), a simple checklist, common troubleshooting steps with fixes, and two short training exercises. Keep language simple for non-technical users and limit the main steps to 8 or fewer. End with success metrics to measure onboarding effectiveness (eg. time to first task, error rate).”

Variants

• Short checklist version: “Create a 10-item quick-start checklist for [ROLE] to start with [TOOL NAME].”
• Detailed version: Add “include exact menu names and sample text to paste into the tool.”

Example (quick)

For Trello onboarding of a Marketing Coordinator the AI might produce: purpose, 6 steps (create board, add lists, create cards with templates, assign members, set due dates), a 5-item checklist, 2 troubleshooting items (card not visible, notification settings), and a 30-minute training task.

Common mistakes & fixes

• Rushed scope — fix: limit to one role and one core workflow.
• Too technical — fix: ask AI to use plain English and add screenshots.
• No testing — fix: always have a new user run the SOP once before publishing.

5-day action plan

1. Day 1: Define scope and map current steps with an SME.
2. Day 2: Run the AI prompt and create draft.
3. Day 3: Review and revise with SME.
4. Day 4: Test with a new user and collect changes.
5. Day 5: Finalize, add screenshots, publish.

Start with one SOP, get it used, then scale. Small, useful wins build momentum — and AI speeds the drafting so you can focus on real-world testing and improvement.

Nice quick win — that single outcome sentence is gold. It gives the headline every interviewer wants. I’ll build on that with a simple, repeatable process and ready-to-use AI prompts so you can turn projects into crisp stories fast.

Why this matters: Interviewers hear tasks. They hire for decisions and results. Your job is to convert project noise into a clear narrative with stake → action → impact.

What you’ll need:

• One-line outcome (your 5-minute headline).
• Project notes: timeline, team, role, baseline metrics.
• Three concrete actions you led and any measurable results.
• A short lesson or what you’d do next time.

Step-by-step (do this now):

1. Frame the problem: 1 sentence with context + risk. (e.g., “Churn at 12% threatened revenue.”)
2. Define your role: 1 sentence — start with “I led…” or “I decided…”
3. List 3 actions you owned — keep them outcome-focused, not tool-heavy.
4. Quantify results: % change, $ saved, time cut. If exact numbers are missing, estimate and label as estimate.
5. Write the lesson: 1 sentence that shows learning and repeatability.
6. Polish with AI: paste your 5-sentence draft into the prompt below to get a 30-second pitch, a STAR answer, and resume bullets.

Example — raw 5-sentence draft:

“Customer onboarding took too long, causing a 12% churn. I led a cross-team redesign as Product Lead. We mapped the process, removed 4 handoffs, and introduced an automated checklist. Churn fell to 7% in 4 months and time-to-value dropped by 30%. I learned to prioritize handoffs that block customer progress.”

Transformed outputs (what to expect):

• 30-second pitch: “I led a redesign of onboarding that cut handoffs, reduced churn from 12% to 7% in 4 months, and sped time-to-value by 30%—by focusing on the touchpoints that blocked customers.”
• STAR answer: Two short paragraphs explaining Situation & Task, then Actions & Results with numbers and the lesson.
• Resume bullets: 2–3 lines with metrics and your specific decision language (“Led, prioritized, removed, automated”).

Common mistakes & fixes:

• Vague metrics — Fix: add percentages or time saved, even conservative estimates.
• Technical detail overload — Fix: translate tech choices into business outcomes.
• Passive language — Fix: use “I” to show ownership.

One-week action plan:

1. Day 1: Create one-line outcomes for 3 projects.
2. Day 2: Draft 5-sentence stories and gather metrics.
3. Day 3: Run the AI prompt for each story (below).
4. Day 4: Edit tone, make cheat-sheet slides.
5. Day 5: Practice 30-second pitches aloud.
6. Day 6: Mock interview and collect feedback.
7. Day 7: Finalize resume bullets and LinkedIn copy.

Copy-paste AI prompt (use as-is):

“I ran a project with this raw draft: [PASTE YOUR 5-SENTENCE DRAFT]. Rewrite into: 1) a 30-second interview pitch, 2) a STAR-format 2-paragraph answer for behavioral questions, and 3) three resume bullets with measurable metrics. Keep language simple for a non-technical audience, emphasize decisions I owned, and limit each item to 150 words or fewer.”

Prompt variants (senior / concise):

• Senior tone: “Same as above, but use a leadership tone highlighting strategy, trade-offs, stakeholder alignment, and long-term impact.”
• Concise: “Create a one-line headline, a 30-second pitch (≤40 words), and two resume bullets (≤20 words each).”

Your next move: pick one project and write that 5-minute headline now. Then paste the five-sentence draft into the prompt and iterate once. Small experiments beat perfect plans.

Quick yes: yes — AI can turn your 2D product photos into realistic 3D renders. But the difference between “good” and “shop-ready” is a simple, repeatable process. Here’s a practical playbook you can use today.

What you’ll need

• Consistent photos (ideally 4–12 angles) on a plain background.
• One or two measurements (height/width or a reference object).
• Basic image editor to crop/remove backgrounds and fix exposure.
• An AI image-to-3D tool or photogrammetry/NeRF option and a 3D viewer.
• 30–90 minutes per product for one pass + quick refinements.

Step-by-step — do this first

1. Choose one product to test (something simple like a ceramic mug or a t-shirt).
2. Take photos: neutral lighting, plain background, at least front, back, two sides, top.
3. Prep images: remove background, normalize exposure, note a measurement.
4. Feed the images and measurement into the AI tool. Ask for texture preservation and PBR materials if available.
5. Open the model in a viewer. If parts are missing, add one or two photos and re-run or touch up in a simple 3D editor.
6. Export: create a studio thumbnail (PNG) and a lightweight 3D file (GLB/USDC) for AR.

Copy-paste AI prompt (use with your tool)

Convert these product photos into a photorealistic 3D model. Input: 6 images (front, back, left, right, top, angled) and product measurements: width 10cm, height 12cm, depth 8cm. Preserve original textures and colors. Produce PBR materials (baseColor, roughness, normal) and export a web-ready GLB under 2MB. Create a studio-render thumbnail (white background, soft 3-point lighting). Optimize geometry for mobile AR without losing visible texture detail. If geometry has holes or missing sides, use photos to reconstruct and report issues.

Prompt variants

• Catalog: “Photorealistic 3D model for product catalog, studio lighting, neutral white background, consistent scale.”
• AR/Interactive: “Lightweight GLB optimized for mobile AR, accurate scale, preserved textures, under 2MB.”
• Marketing hero: “High-res render with enhanced materials and dramatic rim lighting; maintain true color tones.”

Common mistakes & fixes

• Problem: Missing thin parts or transparency. Fix: add side photos and, if needed, mask transparency in a 3D editor.
• Problem: Color shifts. Fix: include a color card in photos and ask tool to preserve baseColor.
• Problem: Heavy geometry or huge file. Fix: request LODs or a mobile-optimized GLB export.

Quick example

Took 8 photos of a ceramic mug, fed them into a NeRF tool with a 9cm height measurement. Result: usable GLB in 45 minutes and a clean thumbnail. Two small glitches on the handle fixed by adding one extra side shot.

Action plan — your next 60 minutes

1. Pick one SKU and take 6 clear photos.
2. Run the AI prompt above in your chosen tool.
3. Inspect the result, add one extra photo if needed, export GLB + thumbnail.

Start small, repeat often. Each run teaches the system and improves speed. Expect the first few to need a tweak — that’s normal. Focus on consistency and you’ll get scalable, realistic 3D product assets fast.

Spot on: one clear success metric per task plus short daily check-ins is the fastest way to see real performance. Let’s add a simple, AI-powered hiring funnel so you can shortlist in days, not weeks.

Big idea: build a repeatable “Role Pack” and a 45-minute micro-trial that AI can score. It keeps you objective, reduces back-and-forth, and makes affordable talent easier to compare.

What you’ll need

• An AI assistant (ChatGPT or similar) and a basic spreadsheet.
• A short application form (or a shared doc) to collect answers and links.
• A video tool for a 60–90 second intro recording.
• Budget for a 1–4 hour paid trial.

Step-by-step (fast, practical)

1. Create a Role Scorecard (10 minutes): List five outcomes and add weights. Example weights: communication 25%, accuracy 35%, speed 20%, following instructions 20%. This becomes your VA Readiness Index.
2. Generate your Role Pack with AI (5 minutes): Use the prompt below to produce a short job ad, three screening questions, a 45-minute micro-trial, a 5-point rubric with weights, a 7-day onboarding checklist, a one-page SOP outline, and two email templates (invite/decline).
3. Build a simple application form (15 minutes): Ask for: answers to the three screening questions, a 60–90 second intro video link, and availability. Add one attention check (e.g., “Put the word ‘maple’ in your subject line”). Expect fewer, higher-quality applicants.
4. Auto-summarize applications with AI (10 minutes per 10 applicants): Paste each applicant’s answers into your AI with your rubric and ask for a one-paragraph verdict plus 1–5 scores. Add those to your spreadsheet. Shortlist the top 3–6.
5. Run a 45-minute micro-trial (same day): Give three bite-sized tasks that mirror day-to-day work. Pay for this time. Keep instructions crisp and measure the result against your rubric.
6. Score with AI (10 minutes): Paste each artifact into AI with the scoring prompt (below). Record the weighted score. Invite the top 2 to a 20-minute interview.
7. Interview to confirm habits (20 minutes): Focus on reliability signals: past routines, how they handle blockers, and timezone overlap. Use your AI-generated script and score 1–5.
8. Onboard with a 7-day ramp (daily 15-minute check-ins): Share your one-page SOP and success metrics. Track outputs, not effort. Decide at day 7 with the same rubric.

Example micro-trial (45 minutes total)

• Email triage (20 min): Provide three sample emails (customer query, scheduling conflict, invoice follow-up). Ask for: subject line, draft reply, label/tag, and urgency (low/med/high). Score clarity, tone, and judgment.
• Data tidy + summary (15 min): Share a small sample table (10–30 rows) with a few duplicates and typos. Ask them to clean it and give a 3-sentence summary (totals, top item, one suggestion). Score accuracy and following instructions.
• Process recap (10 min): 60–90 second video: “What you did, where you got stuck, what you’d improve.” Score communication and ownership.

Copy-paste AI prompt: Role Pack Generator

“You are my hiring co-pilot. Build a complete Role Pack for a Virtual Assistant focused on email, calendar, and simple Google Sheets. Deliver: a) a 120-word job ad (include hours and budget range I can edit), b) three screening questions with one attention-check instruction, c) a 45-minute micro-trial with three tasks (email triage, data tidy + summary, 60–90s video recap) including exact instructions and expected outputs, d) a 5-point scoring rubric with weights (communication 25%, accuracy 35%, speed 20%, following instructions 20%), e) a 7-day onboarding checklist with daily goals and a 15-minute check-in agenda, f) a one-page SOP outline (tools, steps, quality bar, escalation), g) two short email templates (invite to trial, polite decline). Make everything concise, skimmable, and ready to copy-paste.”

Copy-paste AI prompt: Application Summarizer

“Score this candidate against the VA rubric. Input includes their three answers and attention-check result. Return: 1) a one-paragraph verdict, 2) scores 1–5 for communication, accuracy, speed, following instructions, 3) a pass/fail based on weights (pass if weighted score ≥4 and no category <3), 4) top 2 risks, 5) hire/decline recommendation.”

Copy-paste AI prompt: Trial Evaluator

“Evaluate the following trial outputs. Part A: three email drafts with labels and urgency; Part B: cleaned table summary; Part C: video transcript. Score on communication, accuracy, speed (based on timestamps), and following instructions. Use the weights provided, show the math, and end with hire/decline and one coaching tip if hired.”

Insider tricks that save time

• Instruction trap (harmless): Hide a small instruction mid-brief (e.g., file name format). It measures attention to detail without being unfair.
• Transcript-first reviews: Run the video through AI to summarize clarity and confidence before you watch. Then watch only the finalists.
• Timezone sanity check: Ask for two daily overlap windows in their local time and yours; confirm during interview.
• Reference micro-script: Three questions max: “What did they own? How did they handle mistakes? Would you rehire?” Keep it short; score notes 1–5.

Common mistakes and quick fixes

• Overlong trials: Keep it under an hour for screening; save bigger projects for paid extended trials.
• Unweighted scoring: Weight accuracy higher than speed early on. Consistency beats rushing.
• Vague acceptance bar: Set thresholds (e.g., weighted ≥4, nothing <3) before you review to avoid bias.
• One-and-done onboarding: Use daily check-ins for week one, then weekly reviews. Same metrics, less noise.

48-hour action plan

1. Today: Write your five outcomes and weights. Run the Role Pack Generator prompt. Paste outputs into your form and spreadsheet.
2. Next 24 hours: Post the ad. Collect applications. Use the Application Summarizer to shortlist 3–6.
3. Day 2: Run the micro-trial for the shortlist. Score with the Trial Evaluator. Interview top 2. Hire one for a 7-day ramp.

Closing thought: Hire for reliability, test for skill, coach for speed. With a Role Pack, a micro-trial, and a weighted rubric, AI helps you move fast without gambling on fit.

Nice work: your 7‑minute loop is tight and realistic. Locking key terms and aiming for a 35–45% cut keeps meaning while making the copy faster. Let’s add one pro move that saves a full pass most days: diagnose first, then rewrite.

Why this matters: most “Grade 7” misses come from two culprits — a few long sentences and 3–5 jargon terms. A quick diagnosis shows you exactly where to focus, so your rewrite hits target on the first try.

What you’ll need

• Your text in 120–150‑word chunks.
• A short protected list: brand names, prices, deadlines, CTAs.
• Tone tokens: three words that describe your voice (example: “warm, direct, professional”).
• Banned words that feel too corporate (example: “synergize, leverage, optimize”).
• An AI assistant (any chat tool) and a readability check (or ask the AI to report metrics).

Step‑by‑step (diagnose → rewrite → polish)

1. Diagnose (2 minutes): Ask the AI to rate each sentence for grade level, flag jargon, and list the top 5 changes that will drop the grade fastest. You’ll fix the few lines that cause most of the problem.
2. Rewrite with locks (3 minutes): Tell the AI to rewrite to Grade ≈7, keep protected items verbatim, cut 35–45% of words, and use your tone tokens. Ask it to define any uncommon term once, in one short clause.
3. Layer for clarity (optional, 1 minute): Request a two‑layer output: an executive summary at Grade 6–7 and a short details section at Grade 8–9. This keeps simple readers moving and gives experts what they need.
4. Flow stitch (1 minute): Read it aloud. If it feels choppy, ask the AI to join a pair of short sentences where it improves flow, while keeping average sentence length under 15 words.
5. Final check (1 minute): Confirm FK ≈7 and Reading Ease >60. Verify that numbers and CTAs are unchanged. If any nuance was lost, restore it with a brief clause, not a long sentence.

Copy‑paste prompts (robust and ready)

• 1) Diagnose first“Diagnose the text below for a 7th‑grade target. For each sentence, return: a) estimated grade level, b) jargon or hard words, c) the simplest accurate rewrite for that sentence. Then provide: 1) top 5 global changes that will drop the grade fastest, 2) overall metrics (Flesch Reading Ease, FK Grade, average sentence length), 3) any nuance at risk. Text: [paste 120–150 words].”
• 2) Rewrite with locks, tone, and compression“Rewrite the text below to a 7th‑grade reading level (FK ≈7, Reading Ease >60). Keep these protected items verbatim: [Brand, $Price, Deadline, CTA]. Use tone: [warm, direct, professional]. Avoid these words: [synergize, leverage, optimize]. Cut total word count by 35–45% while preserving meaning. Use short sentences (avg <15 words), active voice, and define uncommon terms once in a short clause. Return: 1) simplified text, 2) metrics (Reading Ease, FK Grade, avg sentence length, word count and % reduction), 3) change log (what you cut or rephrased), 4) lost nuance (1 line), 5) three teach‑back questions a 12‑year‑old can answer after reading. Here is the text: [paste your text].”

Worked example (short)

• Original (38 words): “Our integrated platform streamlines operational workflows to optimize resource allocation and drive measurable ROI across departments, enabling leaders to leverage data‑driven insights for continuous improvement and strategic alignment.”
• Diagnosis highlights: Grade ≈ 12. Hard terms: integrated, streamlines, operational workflows, optimize, ROI, leverage, data‑driven, alignment. Long sentence (28+ words).
• Rewrite (22 words): “Our platform makes work easier. It helps teams use time and budget better. Leaders see clear results and make smarter changes.”
• Expected metrics: FK ≈ 6.8–7.2, Reading Ease >60, avg sentence length 7–9 words, ~40% fewer words.

Insider guardrails that boost clarity

• One idea per sentence: Split at “and,” “which,” or “by.”
• Three‑syllable rule: Prefer simpler synonyms when a precise one exists (use “use” not “utilize”).
• Term ladder: Define a complex term once, then use the short form. Example: “ROI (money gained from what you spend).”
• Read to a pause: If you need to breathe, the sentence is too long. Split it.

Common mistakes and quick fixes

• Meaning drift from compression: Fix by setting a range (35–45%) and asking for a change log.
• Tone gets too casual: Fix by naming tone tokens and banning 2–3 words that don’t fit your brand.
• AI edits your prices or CTAs: Fix by tagging protected items and saying “keep verbatim.”
• Choppy rhythm after simplification: Fix with a “flow stitch” pass; keep average sentence length under 15 words.

What to expect

• 1–3 passes to land at Grade ≈7 with Reading Ease >60.
• 35–45% fewer words while keeping facts and CTAs intact.
• Cleaner inboxes: fewer clarifying questions from readers.

5‑day plan (light lift, real gains)

1. Day 1: Make your protected list and tone tokens. Save them as a reusable header for prompts.
2. Day 2: Run the Diagnose prompt on one email and one page. Note the top 5 global changes.
3. Day 3: Run the Rewrite prompt. Restore any missing nuance with one short clause.
4. Day 4: Do a flow stitch, then A/B test simplified vs. original with a small audience.
5. Day 5: Roll out the winner. Save your best prompt + settings as a template.

Closing nudge: Keep the 7‑minute loop. Add the 2‑minute diagnosis up front. You’ll cut a full iteration, protect meaning, and hit Grade 7 with less stress — day after day.