Jeff Bullas

Nice focus — turning research into storyboards is exactly the right way to make findings useful. It moves dry data into decisions people can act on.

Quick promise: I’ll show a simple, repeatable process you can use today to turn research findings into clear visual storyboards using AI — no technical skills required.

What you’ll need

• Raw research: notes, quotes, key stats.
• A text AI (Chat-style) to summarize and craft captions.
• A visual tool (slides, Canva, or an AI image generator) to build frames.
• 5–10 minutes per storyboard frame.

Step-by-step: from research to storyboard

1. Extract the essentials. Paste your research into the text AI. Ask for 3–5 core insights and a 15-word summary.
2. Pick a narrative arc. Use 3–6 frames: Context → Problem → Insight → Evidence → Recommendation → Next step.
3. Write short captions. For each frame, create a 10–20 word caption that answers: What? So what? What now?
4. Generate simple visuals. Use icons, bar/line charts or a single illustrative image per frame. Keep visuals uncluttered.
5. Assemble into slides. One frame per slide. Use large fonts, one visual, one caption.
6. Test quickly. Show to one stakeholder, capture two improvements, iterate.

Copy‑paste AI prompt (use this in your text AI)

“You are a concise research summarizer. Given the following research notes, produce: 1) three core insights each in one sentence; 2) a one‑line problem statement; 3) five storyboard frame captions (10–20 words each) that map to: Context, Problem, Insight, Evidence, Recommendation. Keep language simple for a non-technical audience.”

Visual generation prompt (for an image AI or slide tool)

“Create a clean illustration showing [frame idea e.g., an employee working from home with clock and graph], flat style, minimal colors, high contrast, one focal element, suitable for a slide background.”

Example (remote work study)

• Context: 62% of employees prefer hybrid work.
• Problem: Productivity dips in unstructured home weeks.
• Insight: Short, scheduled collaboration blocks boost output.
• Recommendation: Schedule two focused team days and one async day.

Mistakes & fixes

• Too much text: fix by trimming captions to one sentence.
• Busy visuals: fix by using one chart or icon per frame.
• No narrative flow: reorder slides to tell a problem→solution story.

7‑day action plan

1. Day 1: Gather research and paste into AI.
2. Day 2: Create 3–6 captions and select visuals.
3. Day 3–4: Build slides and refine visuals.
4. Day 5: Test with one stakeholder.
5. Day 6–7: Revise and finalize.

Reminder: Start small, show quickly, then improve. A simple, clear storyboard creates buy‑in far faster than a 40‑page report.

Nice point — wanting photorealistic backgrounds that match your product’s lighting and scale is exactly the right focus. That attention to realism makes the difference between amateur and professional ecommerce images.

Here’s a practical, step-by-step approach you can use today to create photorealistic backgrounds with AI and swap them into your product photos.

What you’ll need

• Consistent product photos (same camera angle, distance, and lighting if possible).
• A simple photo editor (to mask or crop your product) — many free options exist.
• An AI image tool that supports text-to-image and image editing (inpainting/img2img).
• Basic folder to save originals and versions.

Step-by-step process

1. Photograph your product on a plain background (white or neutral). Keep the camera fixed and use one light setup for a batch of shots.
2. Remove the background or create a clean mask around the product. Save a transparent PNG.
3. Decide the scene you want (studio, soft indoor, beach, wood table, etc.). Be specific about time of day and mood.
4. Use an AI image generator to create a matching background. Include lighting, perspective, and depth cues in the prompt.
5. Composite the product PNG over the generated background. Adjust scale, shadow, and color balance so the product looks grounded in the scene.
6. Refine: generate alternate backgrounds, tweak prompts, and test on your storefront to see which converts better.

Copy-paste AI prompt (use as-is; edit product details)

Prompt: Create a photorealistic background for a product photo: a warm indoor wooden table scene at late afternoon golden hour, soft directional light from the left, shallow depth of field, subtle bokeh in the background, neutral warm color grading. Include natural soft shadows and correct perspective for a product placed in the center foreground. High resolution, realistic textures, no people, keep composition simple for overlaying a product PNG.

Example workflow

1. Shot: white background mug photo, camera on tripod, light from left.
2. Mask mug, save png.
3. Generate background with the prompt above, asking for slightly lower brightness to match the mug’s shadow direction.
4. Composite, add a faint shadow under mug, nudge color temperature to match.

Common mistakes & fixes

• Product looks “cut and pasted”: add a soft shadow and slight color grading to match — lower opacity multiply shadow layer works well.
• Lighting mismatch: note the light direction in your prompt (“light from left, soft shadow”) and re-generate.
• Scale feels wrong: use a reference (a plate or hand) in one test scene to get proportions right, then replicate.

Quick 3-step action plan (today)

1. Shoot 10 products with the same setup.
2. Mask them, then generate 3 different background styles using the prompt above.
3. Composite, test on product pages, and keep the top-performing style.

Start small, test what converts, and iterate. Photorealism is achievable quickly when you control the original photo’s lighting and use precise prompts. Do one batch this week and you’ll have a noticeable upgrade to your product images.

Spot on about calibration — the thermometer analogy is perfect. A score only earns trust after it’s checked against real outputs. Let’s turn that idea into a simple, repeatable system your team can run this week.

Quick win: Add a second guardrail beside confidence: a simple Red/Amber/Green “risk lane” the model must assign to its own output. That extra self-check becomes a reliable trigger for human review faster than confidence alone.

What you’ll set up

• One-page guardrail checklist (tone, banned claims, PII rules).
• Two prompt templates: a creator and a checker.
• A tiny “claims library” (approved phrases you can safely reuse).
• A “no-release list” (topics that always require human sign-off).
• A reviewer checklist and a shared log for flagged items.

How to do it — step-by-step

1. Define risk lanes (write this at the top of your checklist):
   • Green: Factual info or how-to with no numbers, no advice, no PII.
   • Amber: Mentions numbers, policy, or third-party claims; cites sources.
   • Red: Legal/medical/financial topics, guarantees, health outcomes, or any PII. Always human review.
2. Create your “claims library” (5–10 reusable, safe phrases). Examples: “According to our policy…”, “Estimated timeframe…”, “We can’t provide professional advice…”, “Results vary…”. This cuts hallucinations and keeps tone consistent.
3. Write a no-release list: medical guidance, investment promises, exact savings/ROI, personal data, unverified statistics, competitor comparisons. These never go live without review.
4. Install the creator prompt (below). It forces the model to: pick a risk lane, cite sources or say “no reliable source found,” flag PII, and include a confidence score.
5. Install the checker prompt (below). Use it as a separate pass on anything Amber or Red to catch claims and tone drift.
6. Run a 25-prompt calibration sprint: include easy, medium, and tricky tasks. Log for each: lane picked, confidence, sources, reviewer decision (approve/fix/reject), reason. Adjust thresholds based on disagreements.
7. Set your simple gate: Red → review required. Amber with confidence < 0.7 → review required. Anything with PII FLAG → review required. Green ≥ 0.7 can publish with spot checks.
8. Hold a 15-minute weekly safety stand-up: review 5 flagged items, update claims library, add one new example to the no-release list.

Copy-paste prompt: Creator

Act as our brand-safe content assistant. Produce the requested draft and then output a short RiskCard. Rules: 1) Use a friendly, professional tone. 2) Do not provide legal, medical, or financial advice—if asked, say: I can’t provide professional advice; please consult a qualified professional. 3) Do not invent facts, dates, statistics, or monetary figures. If unsure, say you are unsure and provide sources or say no reliable source found. 4) Avoid guarantees or outcome claims; use approved phrases from our claims library where relevant. 5) If any personal or sensitive data appears, write FLAG and explain why. 6) At the end, provide: Risk lane (Green/Amber/Red), confidence 0–1, and bullet sources. If lane = Red or confidence < 0.7 or includes FLAG, append HUMAN REVIEW REQUIRED.

Copy-paste prompt: Checker

You are a brand and compliance checker. Review the draft below against our guardrail checklist. Output: 1) List of risky claims or numbers; 2) Whether sources substantiate each claim (yes/no); 3) PII findings (FLAG if any); 4) Tone mismatches (with a suggested fix); 5) Final decision: APPROVE, FIX, or REJECT; 6) If FIX/REJECT, provide an edited paragraph that is safe and on-brand.

Example (what good looks like)

• Task: Write a customer email about a delivery delay and refund options.
• Expected:
  • Friendly, calm tone; no promises beyond policy.
  • Mentions where to find policy: “See our Refunds Policy, section on delays.”
  • No exact compensation amounts unless they’re policy-backed and sourced.
  • RiskCard: Amber, confidence 0.8, sources listed, no PII FLAG. If a number appears, HUMAN REVIEW REQUIRED.

Insider trick: Add a tiny “never words” strip to your prompt components. Examples: guarantee, cure, risk-free, best-in-class, certified, ROI, insider, secret, overnight, safe for all. Any appearance → Red lane. This is a cheap, high-signal filter for legal and reputation risk.

Common mistakes and fast fixes

• Trusting confidence blindly — Pair it with lanes and a checker pass. Calibrate on 25 real prompts before you set gates.
• Prompt bloat — Keep components modular: tone block, safety block, PII block, sources block. Easier to tune and reuse.
• Source holes — Allow “internal policy name/ID” as a valid source; forbid unlabeled stats.
• PII creep — Instruct masking by default (e.g., [Customer First Name], [Order ID]). Only unmask with explicit consent and reviewer approval.
• Channel drift — Lock channel rules: social posts = Green only; ads and emails = Green/Amber; Red never publishes.

One-week action plan

1. Day 1: Finalize risk lanes, no-release list, and the never-words strip.
2. Day 2: Save the Creator and Checker prompts; add the disclaimer to templates.
3. Day 3: Build a 25-prompt calibration set; include edge cases with numbers and sensitive topics.
4. Day 4: Run the sprint, log disagreement between lanes/confidence and reviewer calls.
5. Day 5: Tune thresholds; update the claims library with 5 approved phrases you’ll reuse.
6. Day 6: Train reviewers on the 5-point checker output; run two live tests.
7. Day 7: Go live with Green ≥ 0.7 auto-publish, Amber/Red gated; schedule the weekly safety stand-up.

What to expect: A small slowdown at first, then faster, safer publishing as lanes + checker reduce noise. You’ll see fewer risky phrases, more consistent tone, and clearer sourcing. Keep iterating your claims library — it’s the easiest lever to scale safe, on-brand content.

Simple beats perfect. Start with lanes, the two prompts, and a 25-prompt sprint. You’ll put real guardrails in place without killing speed.

Nice point — yes, giving AI clear constraints and doing a light editorial pass is the secret sauce. That’s where speed meets authenticity.

Why this works

AI is fast at structure: hooks, segment flow, question banks and draft scripts. You add voice, fact-checks and the human follow-ups that make a conversation memorable. Below is a practical playbook you can use right away.

1. What you’ll need
   • Episode topic and single goal (e.g., teach one tactic, spark debate)
   • Audience snapshot (age, experience, interest)
   • Guest bio bullets and hot-button views
   • Preferred tone, episode length and language examples (short sample lines)
2. Step-by-step — do this with AI
   1. Give the brief (topic, goal, audience, guest bullets, tone).
   2. Ask for 2 hooks and a 5-point segment outline with timestamps.
   3. Request three layers of questions: warm-up (3), deep-dive (5), provocative follow-ups (3).
   4. Generate a timed draft script: intro (30s), transitions (20s each), closing (30s).
   5. Run a quick fact-check on any claims; adapt phrasing to the guest’s voice.
   6. Rehearse once, mark natural pauses and ad-libs, then finalize show notes and social blurb.

Quick example — topic: Remote Work Burnout

• One-paragraph hook: “Many professionals love remote work but struggle to switch off. Today we unpack what causes burnout and practical fixes you can try this week.”
• Warm-up questions: 1) “How did you personally realise you were burned out?” 2) “What’s one small daily habit that helped?” 3) “What surprised you about recovery?”
• Deep-dive: 5 evidence-backed questions that probe causes, systems, and measurement.
• Closing: “Top 3 takeaways and one action for listeners this week.”

Mistakes & fixes

• AI gives generic answers — fix: ask for specifics tied to the guest’s bio and one real example.
• Over-reliance on statistics — fix: flag any stat and ask AI to provide a source or rewrite without numbers.
• Script sounds stiff — fix: request a conversational rewrite with contractions and parenthetical notes for ad-libs.

Copy-paste AI prompt (use as a template)

Act as my podcast co-producer. Episode topic: [insert]. Goal: [teach/entertain/debate]. Audience: [brief description]. Guest: [3 bullet points with experience and views]. Tone: [warm/conversational/serious]. Produce: 1) Two 1-sentence hooks, 2) a 5-point timed outline (with minute marks for a 30-minute show), 3) three layers of questions (warm-up 3, deep-dive 5, provocative follow-ups 3), 4) a 30-second intro script and a 30-second closing with three clear takeaways.

Action plan — try this in 30–60 minutes

1. Fill the brief and paste the prompt above into the AI tool.
2. Choose one hook and one outline; generate the layered questions.
3. Do a 15-minute read-through and tweak language to match the guest.
4. Record a short rehearsal and note two places to improvise.

Small experiments drive big learning. Use AI to create drafts quickly, then use your voice and judgement to make the episode truly yours.

Nice call on the bullet-by-bullet approach — that’s the single best shortcut to believable, measurable resume bullets. I’ll build on that with a tight, practical playbook you can use right now.

What you’ll need

• One original resume bullet you want to improve.
• Your job title and the scope (team size, region, project length).
• Any numbers or timeframes you remember (even rough estimates).
• 15–30 minutes per bullet to iterate.

Step-by-step (do this for one bullet at a time)

1. Read the original bullet and ask: why did this matter? who benefited? over what period?
2. Pull any supporting facts: number of people affected, revenue, time saved, frequency, or tools used.
3. Choose a conservative metric style: percent change, ranges, time saved, or $-range.
4. Write three variants: (A) ATS-friendly one-liner, (B) interview-ready line with context, (C) conservative/estimated version.
5. Label estimates clearly when unsure (use “~” or the word “estimated”).
6. Verify quickly (email, spreadsheet, calendar). If you can’t verify, keep numbers conservative and flagged.

Worked example — follow this pattern

• Original: Improved onboarding process for new hires.
• ATS-friendly: Redesigned new-hire onboarding, reducing time-to-productivity by ~25%.
• Interview-ready: Led a cross-functional redesign of the 4-week onboarding program for 30 new hires, shortening time-to-productivity by approximately 20–30% and reducing first-month support tickets by half.
• Conservative (estimated): Redesigned onboarding for ~30 hires, resulting in an estimated 20–30% faster time-to-productivity (est.).

Common mistakes & fixes

• Mistake: Inventing exact numbers. Fix: Use ranges and label them “estimated.”
• Mistake: Leaving out scope. Fix: Add team size, region or project length for context.
• Mistake: Rewriting everything at once. Fix: Batch one bullet per sitting (15–30 minutes).

Quick 5-day action plan

1. Day 1: Pick 3 bullets and gather any evidence (15–30 min each).
2. Day 2: Run the AI prompt for each bullet and review outputs (30–60 min).
3. Day 3: Verify numbers with records or a colleague; adjust phrasing if needed.
4. Day 4: Replace bullets on your resume (ATS-friendly versions) and keep interview-ready lines in your notes.
5. Day 5: Practice a 30–60 second anecdote for each updated bullet.

Copy-paste AI prompt (use as-is)

Rewrite this resume bullet to include measurable outcomes and conservative estimates. Original bullet: “[paste original bullet here]”. Role: [your job title]. Scope: [team size, region, project length]. Known inputs: [any numbers or timeframes you remember]. Produce three outputs: (1) ATS-friendly one-line with a metric, (2) interview-ready one-line with context and outcome, (3) conservative version labelled “estimated” or “approx.” Keep language action-oriented (reduced, increased, delivered, saved). Do not invent precise dollar amounts — use ranges if needed.

What to expect

• One solid sentence you can paste into your resume and two alternates for interviews and sourcing.
• A quick verification step to keep claims honest.
• Biggest ROI from updating 3–4 top bullets first.

Small, honest numbers move the needle. Pick one bullet now and run the prompt — you’ll have a measurable line in minutes.

Spot on — the 5-minute headline test is the right habit. Let’s turn it into a simple “messaging ops” routine so you get consistent pillars, channel-ready copy, and measurable gains without big rewrites.

High-value unlock: build a Pillar-to-Assets pipeline. Each pillar becomes a reusable kit: one promise, one payoff, three proofs, a few objections with counters, and ready-to-paste lines for web, email, and ads. Your team stops reinventing and starts reusing.

What you’ll set up (once)

• Pillar DNA: Promise (6–8 words), Payoff (one line), Proof (3 bullets), Tone (3 adjectives)
• Objection pack: Top 3 objections + one-sentence counters
• Voice rails: Do/Don’t language, banned words, sentence length target
• Asset snippets: hero headline, subhead, CTA, email subject, ad line
• Metrics: conversion, CTR/CPL, time-to-produce, and a “Message Reuse Ratio” (% of new assets that reuse pillar wording)

Step-by-step — 75 minutes to a working, reusable kit

1. Mine real language (15 min): In one doc, paste your brief, 8–12 quotes/support snippets, and 3 competitor headlines. Highlight repeated outcome phrases customers use.
2. Draft three frames (10 min): For each outcome, write four short lines: Problem → Core Benefit → One Proof → Tone. Keep only three frames.
3. Expand with AI (10–15 min): Use the prompt below to create pillar DNA, an objection pack, and channel snippets. Feed it your quotes so it mirrors customer wording.
4. Frontline alignment (15–20 min): Review with sales/CS. Replace any invented phrasing with actual customer words. Confirm objections feel real.
5. Assemble your one-page kit (10–15 min): For each pillar, include the Promise, Payoff, Proof x3, Tone, Objections + counters, and five ready-to-paste snippets (hero, subhead, CTA, email subject, ad line).
6. Deploy and track (ongoing): A/B test one homepage hero and one ad for 2–4 weeks. Track conversion metrics and your Message Reuse Ratio (target 70%+ reuse).

Copy-paste AI prompt (use as-is)

• “You are a senior messaging strategist. Using the product brief, customer quotes, and competitor lines I provide, create exactly 3 messaging pillars. For each pillar output: 1) Promise: a 6–8 word headline using customer language, 2) Payoff: one sentence that completes the promise, 3) Proof: three specific evidence bullets (metrics, social proof, or process), 4) Tone: three adjectives, 5) Objections: the top 3 buyer objections + one-sentence counters. Then generate channel-ready snippets: a) Website hero headline (max 9 words), b) Subhead (one sentence), c) CTA (2–3 words), d) Email subject (max 7 words), e) Short ad line (max 12 words). Finally, include a 5-item consistency checklist for writers/designers and a list of ‘banned words’ to avoid generic fluff. Use the exact words from my quotes wherever possible and flag any invented phrases in brackets.”

Insider trick to keep output tight

• Ask your AI to bold only the words pulled from customer quotes. This makes it obvious where phrasing is real vs. invented so you can edit fast.
• Set a sentence-length rule (max 16 words) for web hero copy to improve scanability and conversions.

Example — FocusFlow (task manager for small teams)

• Pillar 1 Promise: Finish work faster, fewer meetings
• Payoff: Teams ship on schedule because priorities are crystal clear.
• Proof: 25% faster delivery; 90% task clarity; 300+ small teams.
• Tone: direct, encouraging, confident
• Objection + counter: “We’ve tried tools before” → “FocusFlow pairs tasks with priorities so work doesn’t stall.”
• Snippets: Hero: “Ship faster with fewer meetings.” Subhead: “Clear priorities, predictable sprints.” CTA: “Start now.” Email subject: “Ship faster, meet less.” Ad line: “Stop planning loops. Start shipping.”

Quality bar — the D.R.E.A.M. filter

• Distinct: Would a competitor say this? If yes, rewrite.
• Relevant: Does it mirror customer wording? Quote or cut.
• Evidenced: Is there a number, name, or proof?
• Adaptable: Can it flex across web, email, ads without edits?
• Memorable: Nine words or fewer for the promise.

Common mistakes & simple fixes

• Mistake: Pillars drift across channels. Fix: Add the exact hero+subhead+CTA to your CMS/email/ad templates so creators must reuse them.
• Mistake: Vague proof. Fix: Swap generic claims with a metric, named customer, or before/after time saved.
• Mistake: Too many ideas per pillar. Fix: One promise, one payoff, three proofs. That’s it.
• Mistake: Skipping objections. Fix: Capture three objections from sales calls and add counters to every pillar kit.

Operating rhythm (keep it light)

• Weekly (20 min): Add 2–3 new customer quotes. Retire any proof points older than 6 months.
• Biweekly (30 min): Review A/B results. If a pillar underperforms twice, rewrite the Promise, not the Proof.
• Monthly (20 min): Audit Message Reuse Ratio across new assets. Target 70%+.

Bonus micro-prompt for rapid tests

• “Rewrite this pillar’s Promise into 5 website hero headlines (max 9 words), each using one of these angles: speed, simplicity, certainty, savings, social proof. Keep the same proof bullets. Return as a numbered list for A/B testing.”

What to expect

• Week 1: Clear internal language and faster asset creation.
• Weeks 2–4: Uplift in hero-driven metrics and ad CTR; reduced editing cycles.
• Ongoing: A growing library of reusable lines that speed every campaign.

Next move: run the main prompt with your quotes, build the one-page pillar kit, and swap in one hero + one ad today. Log results. Reuse the winning lines everywhere. Consistency compounds.

Nice point — I agree: a consistent lesson template and short 45–60 minute sprints are the heart of a repeatable, low-stress workflow. That keeps you in control while AI speeds the writing up.

Quick win (try in under 5 minutes)

Give AI this simple request: one-sentence course topic + one-line learner profile. Ask for a 5-module outline. You’ll get structure fast and can refine it.

What you’ll need

• One-sentence course topic and one-line learner profile.
• Typical lesson length (30–60 minutes).
• Top 2–3 outcomes per module.
• Text editor and an AI chat tool you prefer.

Step-by-step routine

1. Draft the prompt: Use the copy-paste prompt below to get a module list.
2. Chunk into lessons: Ask the AI to create 4–6 lessons per module using the lesson template (objective, hook, teach, activity, assessment, resources).
3. Pick one lesson: Edit the AI script for your voice and local examples — keep edits small and focused.
4. Pilot quickly: Run the lesson with one learner or record a short video; note 2 improvements and update the script.

Copy-paste prompt — module list

“I’m creating a beginner course. Topic: [one-sentence topic]. Learners: [one-line profile]. Create a compact curriculum with 5 modules, each with 3 clear learning outcomes and a 4–6 lesson breakdown. Make lessons suitable for [lesson length] sessions and keep language simple.”

Copy-paste prompt — lesson script (template)

“Write a single lesson script using this template: Objective (one sentence), Hook (2–3 min), Core teaching (10–20 min) with 3 key points and simple examples, Practice activity (10–20 min) with step-by-step student tasks, Assessment/exit ticket (2–5 min) — one question or quick task, Resources (links or readings). Keep tone friendly and easy for beginners.”

Example (brief)

Topic: Intro to Digital Marketing for Small Business Owners. Module example: Module 1 — Basics of Online Presence. Lessons: 1) Why a website matters, 2) Simple homepage checklist, 3) Intro to social profiles, 4) Quick SEO for beginners.

Mistakes & fixes

• Mistake: Prompt too vague — AI returns bland content. Fix: Add target age, skill level, and session length.
• Mistake: Over-reliance without checking facts. Fix: Always verify examples and stats; add your local context.
• Mistake: Too many outcomes per module. Fix: Limit to top 2–3 outcomes to keep focus.

Action plan (next 7 days)

1. Day 1: Write one-sentence topic & learner line. Run the module-prompt.
2. Day 2–3: Chunk modules into lessons with the lesson-prompt.
3. Day 4: Edit one lesson for voice and clarity.
4. Day 5: Pilot the lesson, collect feedback.
5. Day 6–7: Revise and repeat for next lessons.

Keep it practical: iterate, pilot, and reuse templates. Small, steady steps win — and AI is your fast drafting partner, not the final judge.

Nice callout — the single-spreadsheet + 20-minute weekly routine is exactly the practical foundation most freelancers need. I’ll add a few quick upgrades you can do in the same workflow to make forecasts more actionable and reduce stress fast.

What you’ll need (quick checklist)

• a simple spreadsheet (Google Sheets or Excel)
• last 3–6 months of income (invoices or bank deposits)
• monthly expenses and current bank balance
• a list of proposals out (with a guessed probability: 25%, 50%, 75%)

Step-by-step — setup (45–60 minutes)

1. Paste six months of income into three columns: Month | Total | Bucket (retainer/project/other).
2. Calculate a 3-month moving average. Simple formula example (Google Sheets): =AVERAGE(B2:B4) where B2:B4 are the latest 3 months.
3. Create scenarios: Pessimistic = 90% MA, Likely = 100% MA, Optimistic = 115% MA. Add a column for each scenario.
4. Add proposals as conditional income: multiply value × probability and put in a “Conditional” column.
5. Set a cash buffer target (e.g., 1× monthly expenses) and add a column: Current balance − Buffer to show surplus/shortfall.

Weekly routine (10–20 minutes)

1. Update new payments and scheduled invoices.
2. Refresh moving average and see which scenario you sit in.
3. If pessimistic or below buffer, run one trigger: pitch one lead, chase one invoice, or pause one discretionary spend.

Short worked example

• Last 3 months income: $3,000, $2,200, $2,800 → 3-month MA = $2,667.
• Scenarios: Pessimistic = $2,400 (90%), Likely = $2,667, Optimistic = $3,067 (115%).
• If monthly expenses = $2,500 and current bank = $1,200 → you’re short of 1× buffer and must trigger one outreach or invoice chase.

Common mistakes & fixes

1. Overcomplicating categories — start with 3 buckets. Fix: collapse to retainer/project/other.
2. Ignoring proposals — Fix: always include as conditional income with a probability.
3. No weekly action — Fix: attach a single, non-negotiable trigger to each risk level.

Two copy-paste AI prompts you can use now

Forecast + outreach prompt (paste your 6-month table in place of [PASTE TABLE]):
“I’m a freelance [role]. Here is my last 6 months income table: [PASTE TABLE]. Calculate a 3-month moving average, produce three forecast scenarios (pessimistic 90%, likely 100%, optimistic 115%), list top 3 cash-flow risks, and create one short outreach email to pitch a client and one invoice chase message. Keep both friendly and direct.”

Quick invoice chase only:
“Write a short, polite invoice reminder for invoice #123, amount $X, due date DD/MM, noting that a quick payment helps keep project timelines on track. Keep it under 70 words.”

One-week action plan (do-first mindset)

1. Day 1: Build the sheet and paste 6 months of income (45–60 min).
2. Day 2: Run the AI forecast prompt and copy the outreach/chase templates (10–15 min).
3. Day 3: Update bank balance and scheduled invoices (10 min).
4. Day 4: If flagged, send one outreach email or chase one invoice (10–20 min).
5. Day 5–7: Track results and tweak probabilities or triggers as needed.

Small weekly habits beat complex models. Start with this simple routine, act on the triggers, and you’ll see predictable improvements in 4–8 weeks.