Jeff Bullas

Great point — your clear, repeatable pattern (one idea per slide; title + 3 bullets + one-line speaker note) is exactly the high-leverage move. Here’s a compact, practical add-on that helps non-technical users get from notes to a polished deck fast.

Quick context: Aim for speed and clarity. Use AI for structure and drafts. You keep the teaching: examples, tone, verification.

What you’ll need

• Lesson notes (bullets or a 200–400 word script)
• Any chat-style AI tool you can paste prompts into
• Slide editor (PowerPoint or Google Slides)
• Optional: image library or simple AI image generator

Step-by-step workflow (30–60 minutes)

1. Prep (10 min): Edit notes down to 4–7 key ideas. One idea = one slide.
2. Generate outline (5–10 min): Use the prompt below to get slide titles, 3 bullets, one-sentence speaker note, 2 image keywords, and suggested timing.
3. Quick edit (10–20 min): Swap jargon for everyday language, add a local example, and fact-check one or two claims.
4. Build slides (10–20 min): Paste titles and bullets into slides, add one image per slide, use a clean font and consistent layout.
5. Rehearse (10 min): Read speaker notes aloud; trim to hit your total time.

Copy-paste AI prompt (use as-is)

Convert these lesson notes into a 6-slide presentation for an audience aged 40+. For each slide provide: slide title, 3 concise bullets (6–10 words each), 1 one-sentence speaker note, 2 image keywords, and suggested slide duration in seconds. Keep tone clear, practical, friendly, and suitable for in-person or recorded delivery. Lesson notes: “[PASTE YOUR NOTES HERE]”

Worked example (3-slide extract)

• Slide 1 — Why sleep matters
  • Restores brain and body overnight
  • Boosts memory and focus next day
  • Supports immune and heart health
  • Speaker note: Mention one local stat or simple story.
  • Images: “sleep health”, “resting person”
• Slide 2 — Simple bedtime routine
  • Same bedtime each night
  • Wind down 30 minutes before bed
  • Avoid screens and caffeine late
  • Speaker note: Give one example routine you use.
  • Images: “bedtime routine”, “no screens”

Common mistakes & fixes

• Too much text — fix: move detail to speaker notes and keep 3 bullets only.
• Generic visuals — fix: use the image keywords AI suggested or pick a contextual photo.
• Trusting AI facts blindly — fix: quick fact-check and add a local example.

Action plan (one-session)

1. 10 min: Narrow notes to 5 key ideas.
2. 10 min: Run AI prompt and review output.
3. 25 min: Build slides, add visuals, set font/template.
4. 10 min: Rehearse and trim.

Reminder: Use AI to do the heavy lifting on structure. Your clarity, examples and voice turn a draft into teaching that sticks. Try one lesson today and iterate — small wins scale fast.

Nice call on the 24-hour human review — that little quality gate turns automated notes into reliable action.

Here’s a compact, practical add-on to make the workflow faster and safer for non-technical teams. Small tweaks yield quick wins: fewer follow-ups, clearer owners, and predictable follow-through.

What you’ll need

• Recording (Zoom/Teams cloud or phone with consent)
• Transcript (platform auto-transcript or service)
• Text AI tool (paste-based or integrated)
• One human reviewer and a shared task list or tracker

Step-by-step — do this today

1. Record the meeting and export the transcript immediately.
2. Quick-clean: remove obvious nonsense lines (10–60 seconds).
3. Paste the cleaned transcript into the AI prompt below and run it.
4. Human reviewer (within 24 hours) confirms owners/dates and publishes the task list into your tracker.
5. At the next meeting, measure one KPI: % of action items completed on time.

Copy-paste AI prompt (use as-is)

Paste the meeting transcript after the line below and run this prompt exactly:

TRANSCRIPT:

“You are an executive assistant. Read the transcript below. Output, clearly labeled:
1) Executive summary: 3 short bullets (purpose, outcome, blockers).
2) Numbered action items: for each item give: action (short), owner (assign if unclear), suggested due date (or date range), priority (High/Medium/Low), and a one-line confidence level (High/Medium/Low) explaining if the owner/date came from explicit speech or was inferred.
3) Flag any ambiguous items that need human confirmation.
4) Suggest 3 next steps (who should do what) and provide a one-line risk statement.
5) Provide one ready-to-send email summary (2–3 lines) we can paste to attendees.
Keep everything concise and email-ready.”

Example output (what to expect)

• Executive summary: Align on Q3 priorities; approved budget; blocker: vendor timeline uncertain.
• Action 1: Draft vendor SLA — Owner: Sarah (assigned, inferred) — Due: 2025-07-10 — Priority: High — Confidence: Medium.
• Next steps: Sarah to draft SLA; John to confirm vendor dates; Ops to prepare budget note.

Common mistakes & fixes

• Poor audio → bad transcript. Fix: use headset, ask people to mute, or record a second device.
• AI assigns without context. Fix: require human confirmation within 24 hours for owner/date changes.
• Too many meeting types at once. Fix: pilot on one meeting type for 1–2 weeks, then scale.

1-week action plan

1. Day 1: Run process on one meeting and review output.
2. Day 2: Share AI summary with attendees; collect corrections.
3. Day 3–4: Tweak prompt/transcript cleaning based on errors.
4. Day 5–7: Automate export or paste step if repeatable; track % of confirmed actions within 24 hours.

Start with one meeting this week — paste the transcript into the prompt above and you’ll have a shareable action list in minutes. Small test, big payoff.

Your one-variable approach and the 320/64/32 comparison grid are spot-on. That turns taste into a test. Let me add a few pro moves that make the seed strategy faster, clearer, and easier to decide.

High-value tweaks that change the game

• Seed lock: freeze the baseline (file + settings) so every variation is truly comparable. No accidental nudges.
• Fixed increments: test changes in small, even steps (2%, 4%, 6%). You’ll see a curve, not chaos.
• Luminance lock for colors: keep lightness consistent while changing hue so you’re testing color, not brightness.
• Pixel-fit pass at 32px: align key edges to the pixel grid and use whole-pixel stroke widths. That’s the difference between fuzzy and crisp.
• Branch, don’t restart: only create a new “seed branch” when two options tie on your metrics. Name branches clearly.

What you’ll need

• Your seed file (SVG preferred) and a duplicate-safe workspace folder.
• An editor with grid/pixel preview (any simple vector tool is fine).
• A size sheet (artboards or frames at 320px, 64px, 32px). One dark and one light background.
• A simple tracking note: version name, what changed, pass/fail at 32px, 1-line rationale.

Step-by-step (90-minute sprint)

1. Lock the seed (10 min): clean shapes, center align, convert strokes to known widths (e.g., 2px/3px), and save as seed_v1.svg. Duplicate a “test board” with 320/64/32 slots on light and dark.
2. Set fixed increments (5 min): choose 2–3 variables and define exact steps. Example: letter spacing -2%/-4%/-6%, mark size -10%/-15%/-20%, type weight +1/+2.
3. Run three lanes (35 min):
   • Color (value-locked): swap hues but keep similar lightness. Make 3 versions. Name: seed_color_A/B/C.
   • Spacing/weight: 3 versions with your fixed increments. Name: seed_space_A/B/C.
   • Shape simplify: remove the weakest detail each time or unify corners. 3 versions. Name: seed_shape_A/B/C.
4. Pixel-fit micro pass (15 min): on the 32px artboard, align key horizontals/verticals to whole pixels; set strokes to whole pixels; avoid half-pixel positions. If needed, slightly nudge curves so the silhouette stays clean.
5. Comparison board (10 min): place all 9 versions plus the seed on one sheet at 320/64/32, on light and dark. Add Strong/Neutral/Avoid and a 1-line note.
6. Quick test (10 min): do a 5–10 person forced choice: “Which reads best at a glance?” and a 1-second blur test (zoom out or add slight blur). Record a simple score out of 10.
7. Branch if tied (5 min): if two options tie, create seed_v1A and seed_v1B. Carry only those forward.

Insider checks that save rounds

• Negative-space flip: invert to white-on-black. If it collapses, simplify the interior spaces.
• Three-distance test: phone-at-arm’s-length, across the room, and as a tiny favicon mockup. If it fails any, iterate spacing/weight before color.
• Monochrome first: nail the black/white version, then color is easy.

Copy-paste AI prompts (refined)

• Variation generator (keeps one-variable control)“You are a senior logo designer. Starting from this seed (describe shape, type, and current colors), produce 9 minimal variations that each change only one variable at fixed increments: 3 color swaps with similar lightness (provide hex and estimated L*), 3 spacing/weight tweaks (letter spacing -2%/-4%/-6% or weight +1/+2/+3), 3 shape simplifications (remove least-essential detail each time). For each variation, give: a one-sentence rationale, pass/fail at 32px, and a suggestion for pixel alignment at 32px (which edges to snap). Output a simple table of filenames I can use exactly.”
• Critique and micro-adjustments“Evaluate these logo options [paste short descriptions or image refs]. Score each on legibility at 32px, distinct silhouette, and contrast on light/dark. Recommend precise micro-changes in numbers (e.g., increase letter spacing by 2%, thicken verticals by 1px, reduce inner corner radius by 2px). Suggest which two to carry forward and why.”
• Naming + changelog helper“Create a clear file naming and changelog scheme for my logo iterations with seeds and branches. Use this format: seed version, variable, increment, date. Generate 12 example filenames and a one-line description for each.”

What to expect

• One sprint yields 10 options with clean notes and two obvious frontrunners.
• Most wins come from spacing/weight and shape simplification. Color becomes the finishing touch.
• A 32px pixel-fit pass often makes the result feel sharper without changing the design.

Common mistakes and fast fixes

• Changing multiple variables at once. Fix: keep increments fixed and singular per variation.
• Skipping dark-mode checks. Fix: always test on light and dark backgrounds.
• Ignoring pixel grid at small sizes. Fix: whole-pixel positions and stroke widths at 32px.
• Too many colors too soon. Fix: prove monochrome first; lock luminance when adding color.
• Poor naming. Fix: adopt seed_v1_variable_increment (e.g., seed_v1_space_-04).

2-day action plan

1. Day 1 (90 min): lock the seed, run three lanes (9 variations), pixel-fit at 32px, build the board.
2. Day 2 (60–90 min): quick preference test, branch top 2, apply micro-adjustments, finalize monochrome + color, export small/mono/icon versions.

Bottom line

Decisions get easier when you lock the seed, change one variable in fixed steps, and test at tiny sizes. Do the pixel-fit pass, then pick the frontrunner and ship a real-use version. Progress beats perfection.

Quick start (under 5 minutes): Paste the prompt below into your AI and do one problem now. You’ll get an instant read on whether the difficulty fits you, without a big setup.

Nice build on the easy/target/hard idea. Your plan of a short baseline, simple metrics, and weekly nudges is exactly right. Here’s a small upgrade that makes the AI adapt inside a session, not just between sessions.

Insider trick: the “staircase” dial — a simple 1‑up/1‑down rule used in skill testing. If a problem is a good success (on time, confident), go one step harder; if it’s a miss (slow, unsure, or wrong), go one step easier. It converges to your sweet spot quickly.

What you’ll need:

• A timer (phone is fine).
• A simple tracker (columns: problem #, correct?, time, confidence 1–5, error type).
• Your focus area (one subskill for this week).
• An AI chat you can paste prompts into.

Step-by-step (first session):

1. Choose one subskill (e.g., “fractional coefficients in linear equations”).
2. Start at a middle difficulty. Use the staircase prompt below. Set a light timebox (e.g., 3–5 minutes).
3. Attempt one problem at a time. Log: correct/incorrect, time, confidence, error type (conceptual, calculation, misread).
4. Tell the AI your result using the feedback block. It will auto-adjust difficulty by one step.
5. After 6–8 problems, stop. Ask for a one-paragraph summary of patterns and the next two focus drills.

Copy-paste prompt: Adaptive staircase session

“You are my adaptive practice coach for [topic], focused on [specific subskill]. Run a staircase session: start at difficulty 5 on a 1–10 scale. Give me one problem at a time with this format:

– Difficulty: [1–10]
– Objective: [one line]
– Est time: [minutes]
– Hint (locked): [one hint, but only show it if I type ‘hint’]
– Solution (locked): [full worked solution, but only show if I type ‘solution’]

After each problem I’ll reply with this feedback block:
– Answer: [my answer]
– Correct: [yes/no]
– Time: [mm:ss]
– Confidence: [1–5]
– ErrorType: [conceptual/calculation/misread]
– Felt: [too easy/easy/target/hard/too hard]

Adjust difficulty with a 1‑up/1‑down rule: if Correct = yes AND Time <= Est time AND Confidence >= 3, increase difficulty by 1; if any of those fail, decrease by 1; otherwise keep the same. Keep all problems on [specific subskill]. Every 4 problems, summarize patterns in one short paragraph and refine the next objective. Stop after 8 problems and give me a 5-bullet progress report and what to practice next.”

Fast single-problem version (if you only have 3 minutes)

“Give me one ‘target’ problem in [topic] at difficulty 5 with: a one-line objective, one locked hint, and a worked solution I can reveal on request. Estimate a fair time limit. After I answer, ask me for: correct/incorrect, time, confidence 1–5, and whether it felt easy/target/hard. Suggest the next problem one notch up or down based on that.”

What to expect:

• Within the first 6–8 problems the difficulty should settle near your “just challenging” level.
• Your error pattern will become obvious (you’ll see repeats). That tells you exactly what to drill next.
• Future sessions start closer to the right level because the AI remembers your last settled difficulty and error types.

Example (algebra refresh):

• You set subskill: fractional coefficients.
• AI serves Difficulty 5, objective: “Solve linear equations with fractional coefficients.” Est time: 4 min.
• You solve in 3:20, correct, confidence 3/5 → AI moves to Difficulty 6.
• At Difficulty 6 you misread a negative sign, wrong in 5:10, confidence 2 → AI drops to Difficulty 5 and narrows the objective to “careful distribution with negatives.”
• By problem 6 you’re stable at Difficulty 5–6. The AI suggests a micro-drill: “2 minutes, practice distributing a negative across fractions, 3 items.”

Mistakes to avoid (and fixes):

• Jumping difficulty too fast — Fix: change by one step at a time via the staircase; resist big jumps.
• Too many topics at once — Fix: pick one subskill per week; rotate next week.
• Reading solutions before trying — Fix: keep hint/solution locked; attempt first.
• No timer — Fix: light timebox creates a realistic pace signal for the AI.
• Vague feedback — Fix: always send the feedback block; it’s the fuel for adaptation.

Upgrade your tracker (simple but powerful)

• Add a column: “Why I missed it (one sentence).” You’ll spot patterns faster.
• Tag when a problem moves from hard → target → easy. That’s progress you can feel.
• Weekly note: “Next nudge” (one sentence). Keeps momentum without overwhelm.

7-day action plan:

1. Day 1: Pick one subskill. Run the fast single-problem prompt to gauge level.
2. Day 2: Do a full 8-problem staircase session. Log results.
3. Day 3: Review your two most common error types. Ask the AI for a 5-minute micro-drill on just those.
4. Day 4: Staircase session (6 problems). Stop early if you stabilize.
5. Day 5: Light review: 3 “easy” items to build fluency, then 1 “target.”
6. Day 6: Staircase session (8 problems). Compare average time to Day 2.
7. Day 7: Retest two baseline-style items. Ask the AI for a one-paragraph progress summary and next week’s subskill.

Bottom line: Yes—AI can tailor practice tightly to your level, but it needs two things from you: small, honest signals (correct/time/confidence) and steady, one-notch adjustments. Use the staircase prompt, keep hints locked, and track the trend. The sweet spot finds you faster than you think.

Good point — turning lesson notes into a slide deck is one of the highest-leverage tasks for teachers and trainers. It saves time and focuses your message.

Why this works: AI handles structure and first drafts. You keep the human touch: clarity, context and personality. Quick wins are possible in 30–90 minutes for a short lesson.

What you’ll need

• Clear lesson notes (bullet points or a short script).
• An AI text tool (Chat-style model) you can paste prompts into.
• A slide editor (PowerPoint, Google Slides or similar).
• Optional: simple image library or AI image tool for visuals.

Step-by-step workflow

1. Prep your notes: reduce to 3–7 key ideas. One idea = one slide.
2. Ask AI for a slide outline: get slide titles, 2–4 bullets each, and one-sentence speaker notes.
3. Refine content: edit bullets to match your voice and accuracy.
4. Add visuals: ask AI for image keywords per slide, then add icons/images.
5. Build slides: paste titles and bullets into your slide editor, add visuals and consistent fonts.
6. Rehearse: read speaker notes aloud, trim text and time each slide.

Do / Do not checklist

• Do keep one idea per slide.
• Do use 3–5 bullets, 6–8 words each.
• Do add a one-line speaker note to guide delivery.
• Do not crowd slides with full paragraphs.
• Do not skip fact-checking or localizing examples.

Copy-paste AI prompt (use as-is)

Convert the following lesson notes into a 6-slide presentation. For each slide, provide: a slide title, 3 concise bullet points (6–10 words each), a one-sentence speaker note, and 2 suggested image keywords. Keep language simple and friendly for an audience over 40. Lesson notes: “Why sleep matters; sleep stages; tips for better sleep; common sleep disruptions; when to seek help; quick bedtime routine.”

Worked example (short)

• Slide 1 — Why sleep matters
  • Restores body and mind
  • Improves memory and focus
  • Supports immune health
  • Speaker note: Briefly explain benefits and a quick stat.
  • Images: “sleep health”, “resting person”
• Slide 2 — Sleep stages
  • NREM: light to deep sleep
  • REM: dreaming and memory
  • Cycle repeats ~90 minutes
  • Speaker note: Use a simple analogy — ladder of sleep.
  • Images: “sleep cycle diagram”, “brain at night”

Mistakes & fixes

• Too much text — fix: cut bullets to 3 short phrases and move details to speaker notes.
• Generic images — fix: use image keywords matched to slide topic.
• AI errors or vagueness — fix: verify facts and add local examples.

Action plan (30–90 minute session)

1. 10 minutes: edit notes to 5–7 key points.
2. 10 minutes: run the AI prompt and get the outline.
3. 20–40 minutes: paste into slides, add images, adjust design.
4. 10 minutes: rehearse and trim.

Reminder: Aim for clarity, not perfection. Use AI to accelerate structure; your voice and examples make the deck memorable.

Nice work — quick refinement: include at least one sensitivity reader in addition to your editor when identity or culture is involved. A single reviewer is fine for routine edits, but diverse perspectives catch subtle harms.

Here’s a simple, repeatable approach I use: fast, human‑centered, and practical.

What you’ll need

• One scene or 300–500 words (start small).
• A one‑page (or one‑paragraph) style cheat sheet: preferred neutral pronouns, job‑title swaps, and plot‑critical identity notes.
• An AI editor (chat model) and two human reviewers where possible: an editor and a sensitivity reader.
• A versioning habit: save each pass as v1, v2, v3.

Step-by-step (do this now)

1. Choose one scene and note must‑keep elements: tone, key beats, identity details.
2. Run the AI with a clear prompt (see copy‑paste prompt below) asking for a gender‑neutral rewrite and a short changelog of swaps.
3. Do a quick read for voice and subtext — mark any lines that feel flattened or ambiguous.
4. Make two precise human edits: restore tone, clarify references, or preserve identity where plot‑relevant.
5. Send to your editor and a sensitivity reader; collect one round of feedback and apply targeted fixes.
6. Run a final consistency check for pronouns and names across the scene; save as the next version.

Example

Original: “The hostess waved him over and laughed about his costume.”

Neutral rewrite: “The host waved them over and laughed about the costume.”

Why it works: replaces gendered role and pronoun while keeping action and tone. If relationship or gender is plot‑relevant, restore with a note like: “(retain if identity matters to the scene).”

Common mistakes & fixes

• Over‑neutralizing: Removes character texture. Fix: preserve unique traits that matter to story.
• Pronoun drift: Mixed he/they. Fix: run a consistency pass and ask AI to list all pronouns used.
• Erasing identity: Removes plot‑relevant culture. Fix: flag those lines for sensitivity review before changing.

Copy‑paste AI prompt (use as‑is)

Rewrite the following scene to be inclusive and gender‑neutral while preserving tone, character intent, and all plot‑critical details. Replace gendered job titles and pronouns with neutral alternatives where appropriate. For any line that depends on gender or cultural context, provide two alternate phrasings and highlight it for human review. At the end, output a short changelog listing each pronoun/title swap and one sentence explaining why it was changed. Also list any lines that should be checked by a sensitivity reader.

What to expect

1. AI delivers a usable first draft in minutes (often 60–80% ready).
2. Plan 15–60 minutes of human review per scene depending on sensitivity.
3. Iterate quickly: small batch → review → apply fixes → repeat.

Action plan for today:

1. Pick one scene (300–500 words).
2. Use the prompt above with your AI tool.
3. Make two human edits, then ask an editor and a sensitivity reader for one quick pass.

Reminder: AI speeds the draft — your judgment protects the story. Use AI to do first drafts, humans to make final calls.

Quick win: You can use AI to build a practical monthly content calendar in under an hour — without being technical. Let me show you a simple, repeatable way.

Why this works: A monthly calendar gives focus, reduces decision fatigue, and helps you reuse content across platforms. AI speeds up ideation, outlines, captions and scheduling notes so you can spend more time creating and less time planning.

What you’ll need

• A clear objective (brand awareness, leads, email signups).
• 3–5 content pillars (topics your audience cares about).
• An AI tool (chat assistant like ChatGPT or similar).
• A calendar tool or spreadsheet to record dates and assets.
• Basic content formats: blog post, short video, image post, email.

Step-by-step (do this once per month)

1. Set your goal and audience — one sentence: Who are you helping and what action do you want?
2. Audit and choose pillars — list top 3–5 themes your audience needs this month.
3. Generate weekly themes — assign each week a pillar or mini-series.
4. Ask AI for topic ideas — get 3–5 post ideas per week from the AI.
5. Turn ideas into outlines and captions — use AI to write short outlines, a 300-word blog draft, and 3 caption variations per post.
6. Create a schedule — pick dates, platforms, asset type, and CTAs in your calendar.
7. Repurpose — turn a blog into 3 social posts, one short video script, and one email snippet.

Practical example (small business marketing)

• Week 1: Pillar = Lead generation. Post ideas: “5 low-cost lead magnets”, Instagram tip, 500-word blog.
• Week 2: Pillar = Email marketing. Post ideas: “Subject lines that work”, short video script, newsletter draft.
• Week 3: Pillar = Content repurposing. Post ideas: “Turn one blog into 5 posts”, caption set.
• Week 4: Pillar = Results & offers. Post ideas: Case study, client quote image, offer email.

Common mistakes & fixes

• Mistake: Too many pillars. Fix: Stick to 3–5.
• Mistake: No repurposing. Fix: Always outline 2–3 formats per idea.
• Mistake: Vague prompts to AI. Fix: Give context, audience, tone, and call-to-action.

Copy-paste AI prompt

Act as a helpful content strategist. My audience: small business owners over 40 who want practical marketing tips. My monthly goal: generate leads to an email list. Create a 4-week content calendar with one weekly theme, 3 post ideas per week (blog, short video, social post), one email idea, suggested post titles/captions, and a short CTA for each item. Keep tone friendly and simple.

7-day action plan

1. Day 1: Define goal and pillars.
2. Day 2: Use the AI prompt to generate topics.
3. Day 3: Create outlines and captions with AI.
4. Day 4: Schedule into your calendar and assign creation days.
5. Day 5–7: Batch-create one blog and 3 social assets; repurpose into email.

Reminder: Start small, measure what works, then repeat. The calendar is your tool—use AI to speed execution, not to replace your voice.

Great point — focusing on beginners is exactly where the biggest quick wins are. AI can streamline your grant and accelerator applications without replacing the human judgment that funders want to see.

Here’s a practical, do-first guide to get you started today.

What you’ll need

• Clear project notes: goal, beneficiaries, timeline, budget headline, KPIs.
• The grant/accelerator guidelines and scoring criteria.
• An AI writing tool (chat interface) and a text editor for final polish.
• Time for 2–3 review iterations with a colleague or mentor.

Quick checklist — Do / Do not

• Do: Keep input factual and concise; feed the AI the guidelines and scoring criteria.
• Do: Use AI to draft, then edit for voice and compliance.
• Do not: Paste confidential data or rely on AI for budget numbers without checking.
• Do not: Submit AI text verbatim without human review.

Step-by-step

1. Collect: one-page project summary and the funder’s questions/criteria.
2. Prompt: Give the AI the summary + exact question to answer. Ask for a 200–300 word response, with headings.
3. Iterate: Ask the AI to tighten, simplify, or map to scoring language (e.g., “aligns to Objective A, B, C”).
4. Polish: Human-edit for tone, remove generic phrases, add local proof points or numbers.
5. Check: Verify compliance (word limits, attachments, budget math).

Copy-paste AI prompt (use as-is)

“You are an expert grant writer. Using the information below, write a 250-word executive summary that answers the question: ‘Describe the project and its expected impact.’ Use clear plain language, include one measurable outcome and one sentence on sustainability. Project info: [PASTE YOUR ONE-PAGE SUMMARY HERE]. Funders care about: [PASTE 2–3 KEY CRITERIA].”

Worked example

Raw note: “Teach digital skills to 200 seniors in 12 months. Need $30k for trainers and laptops. Partner: local library.”

AI draft (edited): “We will deliver a 12-month digital skills program for 200 seniors through weekly classes at the local library. Expected outcome: 80% of participants will report improved online confidence and complete a basic digital task assessment. Budget: $30,000 for trainers and equipment. Sustainability: training local volunteers to continue classes after year one.”

Common mistakes & fixes

• Mistake: Vague outcomes. Fix: Use measurable targets (numbers, percentages, dates).
• Mistake: Over-reliance on AI phrasing. Fix: Add local anecdotes or specific partner names.
• Mistake: Ignoring guidelines. Fix: Map each answer to scoring criteria before submitting.

48-hour action plan

1. Day 1 morning: Draft one-page summary and paste into the AI prompt above.
2. Day 1 afternoon: Edit AI output, align to scoring criteria.
3. Day 2: Peer review + finalize budget and compliance checks.

Tip: Start with one question and win it. Build momentum question by question. AI speeds drafting — your judgment wins the funding.