Jeff Bullas

Quick win: In 30–45 minutes you can teach an AI your brand voice and walk away with a one-page style guide plus three ready-to-use templates that actually sound like you.

Why this works: AI helps you compress decisions. You give it a few real samples and three clear attributes, then you nudge and refine. The goal is consistency and speed—not perfection on the first pass.

What you’ll need:

• 3–5 short writing samples (one-paragraph email, a product blurb, one social post).
• 3 single-word brand attributes (e.g., warm, clear, practical).
• 20–45 minutes with an AI chat tool and a doc or notes app to save the guide.

Step-by-step:

1. Collect samples: copy-paste each sample into the chat or attach them if the tool supports uploads.
2. State the attributes: list your three words and any words to avoid (jargon, hype, canned phrases).
3. Ask the AI for a one-page guide: include a 2-sentence voice summary, dos/don’ts, 3 channel samples, and 3 short templates.
4. Review and mark: highlight lines that don’t feel like you and ask for a revision focused on those items.
5. Test & iterate: use one template in real outreach, gather quick feedback, then update monthly.

Practical example (one-page output):

• Voice: Warm and practical — explains benefits simply, with clear next steps.
• Dos: short sentences, name the benefit, end with a clear next step.
• Don’ts: avoid hype, buzzwords, long paragraphs.
• Templates:
  • Email subject + body: “Quick question about [product]” — Hi [Name], I saw you were checking out [product]. One simple way it helps is [benefit]. Have 10 minutes this week to chat?
  • Product blurb: “[Product] saves small teams X hours a week by [how]. No setup fuss.”
  • Social post: “Three quick ways to [solve problem]. Which one will you try?”

Common mistakes & fixes:

• Too generic: fix by asking the AI to replace vague words with specific examples or numbers.
• Overly formal or casual: tell the AI to match a specific sample (paste the sentence) and mimic its rhythm.
• Too long templates: request a maximum word count (e.g., 20–40 words).

Action plan — do this now:

1. Grab three short samples and pick three attributes.
2. Open your AI chat and paste the copy-paste prompt below.
3. Review the guide, pick one template, and use it today.

Copy-paste prompt (use as-is):

“I will paste 3 short writing examples after this. My desired brand attributes are: [attribute 1], [attribute 2], [attribute 3]. Based on those examples, create a one-page brand voice and style guide. Include: a 2-sentence voice summary, 6 clear dos and 6 don’ts, three short channel samples (email subject + 1-sentence body; product blurb of 20–35 words; social post of 15–25 words), and 3 reusable templates. Keep language warm and practical. Avoid the following words/phrases: [list banned phrases]. If any line sounds generic, substitute a specific example or statistic. Output in simple bullet points and short sentences.”

Small reminder: aim for one useful page you’ll actually use. Iterate once after real feedback and your voice will feel natural and consistent.

Make recruiters find you, not just notice you. Small profile changes—keyword-focused headline, achievement-led About, and clear role targets—bring disproportionate results. Use AI to do the heavy lifting quickly.

What you’ll need

• Your current LinkedIn profile text (copy About, Headline, and one or two Experience bullets).
• 3 target job titles or roles you want to attract.
• 10–20 minutes and access to an AI chat (ChatGPT or similar).

Step-by-step

1. Gather your profile text and the 3 roles you want.
2. Run an AI profile-audit prompt (copy-paste below). Ask for quick wins, an optimized headline, About section, experience bullets, and recruiter search terms.
3. Apply the AI’s headline and About edits. Keep headline keyword-rich and role-focused, About achievement-led and concise.
4. Update 3–5 experience bullets using the AI’s achievement bullets—lead with outcomes and numbers where possible.
5. Search your updated headline and skills in LinkedIn’s search bar to check how you rank for your target roles; tweak keywords until they match common job postings.

Copy-paste AI prompt (use as-is)

You are an expert LinkedIn profile optimizer who writes for recruiters and hiring managers. Here is my profile: [PASTE YOUR HEADLINE, ABOUT, AND 1–2 EXPERIENCE BULLETS]. My target roles are: [ROLE 1], [ROLE 2], [ROLE 3]. Please provide: 1) Top 5 quick wins I can implement in 15 minutes; 2) An optimized headline (under 220 characters) with keywords; 3) A concise About section (2 short paragraphs) focused on impact and recruiter search terms; 4) Six achievement-focused bullets for my current role including metrics where possible; 5) A list of 12 recruiter-friendly keywords / skills to add to my profile. Use action-first language, warm professional tone, and output headings so I can copy each section easily.

Prompt variants

• Executive: Add “C-Suite” and “board-level” keyword focus and strategic outcomes (revenue, growth, M&A).
• Technical individual contributor: Emphasize specific tech stacks, certifications, and measurable delivery metrics.
• Career change: Ask the AI to translate past achievements into transferable outcomes relevant to the new field.

Example (before → after)

Before headline: “Marketing Manager open to opportunities.”

After headline: “Digital Marketing Leader • Demand Gen & Content Strategy • 3x Lead Growth | B2B SaaS”

Common mistakes & fixes

• Too vague: Replace “Open to opportunities” with target roles and key results.
• No numbers: Add metrics (%, $ saved, leads generated).
• Keyword mismatch: Mirror language from 3 target job postings so recruiter searches match.

Action plan — do this today

1. Run the copy-paste prompt with your profile (10 minutes).
2. Update headline and About (10 minutes).
3. Refresh 3 experience bullets and add 10 keywords (15–20 minutes).

Closing reminder

Small, tested changes beat big rewrites. Iterate weekly—monitor recruiter views and messages—and keep your profile aligned with the job titles you want. Use the prompt above whenever you target a new role.

Try this now (5 minutes): Paste the prompt below into your AI, fill the brackets, and you’ll get one week of platform-ready posts from a single idea.

Copy‑paste prompt:

Use my voice to create a 1‑week social plan from one idea. Voice: [confident, warm, no jargon]. Audience: [describe]. Pillars: [3–4 pillars]. Platforms: LinkedIn (120–180 words), Instagram (40–80 words), X (max 220 characters), Email (3–4 sentences). Idea: [paste 3–5 bullets or a 400–800 word draft]. Output per day (Mon/Wed/Fri): 1 headline, a platform‑tailored caption for each platform, one 30–60s video script (hook, 3 beats, CTA), and 2 CTA options per platform. Include a simple visual idea for IG/FB and 3 X one‑liners. Keep it punchy, human, and on‑brand.

Why this works: You’re scheduling “slots,” not chasing inspiration. Your pillars repeat, your voice stays consistent, and AI does the slicing so you can focus on quality.

What you’ll need

• Platforms and a conservative cadence (start with 2 posts/week per platform).
• 3–4 content pillars you’ll stick to for 8 weeks (e.g., Insight, Tip, Story/Case, Offer).
• A simple scheduler or calendar (any tool you already use).
• One 60–90 minute weekly block to create a “master asset.”
• A tiny KPI sheet to track: engagement rate, audience growth, and one conversion.

Step‑by‑step: the evergreen + timely rhythm

1. Set your slots. Map pillars to days. Example: Mon=Insight, Wed=Tip, Fri=Story. Leave 1 extra “timely” slot per week open for news, wins, or offers.
2. Create one master asset per week. 400–800 words or a 5–7 minute voice note/video. Pick evergreen topics you can reuse for months.
3. Slice with AI. Use the prompt above to turn that asset into multi‑platform posts, one short video script, and X one‑liners.
4. Tailor your CTAs. Decide one conversion per platform and repeat it. Example: LinkedIn = comments, IG = saves, X = replies, Email = click/answer a question.
5. Schedule a buffer. Always keep 5–7 posts scheduled ahead. If your buffer drops below 3, halve posting for a week and rebuild.
6. Review weekly (15 minutes). Note which pillar and format performed best. Next week, give the winner one extra slot.

Insider template: the 2×2 Message Matrix

• Short + Teach: one micro‑tip or checklist (great for X/IG).
• Short + Proof: stat, quote, or mini case (great hook on LinkedIn/X).
• Long + Teach: step‑by‑step post (LinkedIn/email anchor).
• Long + Proof: story with numbers (case study you can clip into shorts).

Rotate these across your pillar days to keep variety without inventing new ideas.

Quality bar (use this checklist before you schedule):

• Hook in line 1 (question, bold claim, or stat).
• One clear takeaway (teach, bust a myth, or show a checklist).
• One tiny example (numbers or before/after).
• One CTA aligned to the platform.

Micro example (1 topic → 1 week):

• Master idea: “Reduce no‑shows with a 3‑bullet confirmation message.”
• Mon – Insight (LinkedIn): Hook: “Your calendar isn’t leaking time. Your confirmations are.” 150 words teaching the 3 bullets + question CTA. IG: 60‑word caption + quote card. X: “No‑shows drop when confirmations say: time, benefit, reply‑to‑confirm.”
• Wed – Tip: 45s video script: hook + 3 beats (time, benefit, reply). X thread with 3 bullets. IG caption invites saves.
• Fri – Story: Mini case: “From 38% to 19% no‑shows in 3 weeks” with one number and one lesson. Email: 3–4 sentences + soft CTA.

Common mistakes and quick fixes

• Inconsistent voice: Run a one‑time “Voice Guide” prompt and reuse it in every request.
• Overwriting: Add hard caps: LinkedIn 180 words, IG 80 words, X 220 characters. Ask AI for 3 hooks and pick the tightest.
• Platform mismatch: Don’t paste the same caption everywhere. Keep the idea, change the wrapper.
• CTA confusion: One conversion per platform for 4 weeks. Then review and adjust.
• Running out of ideas: Build a “hook bank” of 30 hooks per pillar. Reuse and remix; your audience won’t see them all the first time.

Copy‑paste prompts (save these)

• Voice & Pillars Guide: Act as my social editor. Ask me 5 questions to capture my tone, audience, phrases to use/avoid, expertise, and offers. Then produce a 120‑word Voice Summary and list 3–4 content pillars, each with 2 example angles and 3 sample hooks.
• Repurpose Remix (45‑day cooldown): Here’s a past post: [paste]. Remix it for the same platform with a new angle: [inform/showcase/prompt/myth‑bust/checklist]. Keep the core takeaway, change the hook and examples, and keep length caps: LI 180w, IG 80w, X 220 chars.

14‑day action plan

1. Day 1: Run the Voice & Pillars Guide prompt. Approve it.
2. Day 2: Map your 2‑post/week evergreen slots + one timely slot.
3. Day 3: Create your first master asset (60–90 minutes).
4. Day 4: Run the 1‑Week plan prompt (top of this message). Schedule 5–7 posts.
5. Day 5: Record 2–3 short videos from the scripts. Schedule.
6. Day 6: Build your hook bank (10 hooks per pillar to start).
7. Day 7: Review KPIs and keep winners. Rest.
8. Day 8–14: Repeat with a new master asset. Protect your buffer. If life gets busy, post only evergreen slots.

What to expect

• Weeks 1–2: Slight edits as you lock your tone and CTA.
• Weeks 3–4: 40–60% faster creation, a 5–7 post buffer, and clear pillar winners.
• Beyond: Scale the winners; keep cadence only as fast as your buffer allows.

Last thought: Slots first. Master asset weekly. Strict repurposing rules. That’s how consistency compounds without stealing your time.

Nice point — the AI-first workflow (extract, map, draft, validate) is the fastest way to get readable guides in front of users.

Here’s a practical, do-first addition: simple templates, small automation steps, and a testing loop so you actually measure improvement quickly.

What you’ll need

• Source policy files (PDF/Word). Break long docs into sections.
• Audience personas (role, example tasks).
• Access to an LLM or enterprise AI with document handling.
• Output templates (Guide card, Checklist, FAQ, SME flags).
• One SME and two pilot users for quick validation.

Step-by-step — run this workflow

1. Ingest: Split the policy into 500–1,000 word chunks. Feed each chunk to the LLM with a request to extract obligations, deadlines, exceptions, and required evidence.
2. Map: For each obligation, create a table mapping obligation → role → exact action → frequency.
3. Draft: Use the template to create a 1-minute checklist, two examples (right/wrong), and a short FAQ.
4. Flag: Automate flags for legal nuance (words like “must”, “shall”, “unless”) and send those snippets to SME only.
5. Pilot: Give the guide to 2 users to follow while you time task and collect questions.
6. Publish & Monitor: Add version, review date, and track time-to-task, support tickets, and comprehension quiz score.

Practical example

Policy line: “Employees must encrypt customer data before storage.”

• Guide: Purpose — Protect customer data. Checklist — 1) Open secure storage tool; 2) Use Encrypt button; 3) Confirm file shows lock icon; 4) Upload to approved location. Example correct: file saved with lock icon. Incorrect: file uploaded without encrypting.

Common mistakes & fixes

• Overly technical drafts — fix: enforce 3-step checklist and a plain-English readability check with a pilot user.
• Missing legal nuance — fix: auto-flag keywords and require SME sign-off for flagged items.
• Stale content — fix: add review-date metadata and calendar reminders.

Copy-paste AI prompt (use as-is)

“You are an expert compliance writer. Convert the following policy text into a user-friendly guide for [ROLE] with: 1) one-sentence purpose; 2) a 3–5 step action checklist in plain English; 3) two short examples (correct vs incorrect); 4) a 2–3 question FAQ that clears up confusion. Flag any sentences containing the words: must, shall, required, unless for SME review. Return output as clear sections labeled Purpose, Checklist, Examples, FAQ.”

Prompt variants (quick wins)

• Quick card: “Turn this into a 40-word intranet card for [ROLE] with a 3-step checklist and one example.”
• JSON for tooling: “Return JSON with keys: purpose, checklist[], examples[{correct,incorrect}], faq[]”

1-week action plan

1. Day 1: Pick 1 policy and identify roles.
2. Day 2: Run the main prompt; get a draft.
3. Day 3: SME review flagged items.
4. Day 4: Pilot with 2 users; record time and questions.
5. Day 5: Publish guide with version and review date; add metrics to dashboard.

Quick closing reminder

Start small, measure one thing (time-to-task or tickets), and iterate. AI speeds drafting — the SME and pilot users make it safe and useful.

Spot on: (Impact × Confidence) ÷ Effort is the cleanest way to decide fast. Here’s how AI makes it even better — by turning messy notes into a ranked, evidence-backed short list and a ready-to-run test in under an hour.

Where AI helps most

• Tidy and cluster a long list into clear, non-duplicated insights.
• Add evidence from what you paste (reviews, emails, sales notes) to raise or lower Confidence.
• Estimate effort by breaking work into steps and flagging hidden unknowns.
• Rank in bands (Now / Next / Later) so you stop overthinking tiny score differences.
• Auto-draft a micro-test for the top 1–2 ideas with metric, threshold, and a one-week plan.

What you’ll need

• 5–15 insights (one sentence each)
• Any quick proof you have: snippets from emails, reviews, support tickets, analytics notes
• A simple spreadsheet or notebook
• An AI assistant you can paste text into

Step-by-step: a 45–60 minute AI prioritization sprint

1. Normalize and de-duplicate (10 min): Paste your raw list and ask AI to merge duplicates, fix wording, and group by theme. Expect 5–10 crisp, unique insights.
2. Score with evidence (15 min): Paste relevant snippets. Ask AI to assign Impact, Effort, Confidence (1–5) with one-sentence justifications, apply an Unknowns Multiplier (2×) if effort is unclear, and compute Priority = (Impact × Confidence) ÷ Effort.
3. Band, don’t nitpick (5 min): Have AI label top 3 as Now, next 5 as Next, the rest Later. This avoids debating 8.2 vs 8.4.
4. Draft the micro-test (15 min): For the top item, AI writes a one-week test brief: one change, one audience, one metric, a baseline, a pass/fail threshold, and a short checklist.
5. Calendar and commit (5 min): Put the 7-day plan on your calendar. On day 8, review results and rescore with the new evidence.

Insider trick: Ask AI for an “Evidence Pack” (quotes from your pasted data that support or challenge each insight). You’ll feel more confident setting the Confidence score and faster saying no to weak ideas.

Mini example (before → after)

• Before: “Add a purchase reminder email.”
• AI’s quick Effort Map: draft copy (20 min), segment buyers-without-checkout (30 min), set send rule (15 min), QA (15 min). Unknowns: ESP access? compliance check? → apply 1.5–2× if unknown.
• Scores: Impact 4 (abandoned carts common), Effort 1–2 (simple steps), Confidence 5 (matches reviews: “I forgot to finish”). Priority ≈ high → Run first.

Mistakes and easy fixes

• Hallucinated numbers: AI invents data. Fix: Only let it use the evidence you paste; ask it to quote sources verbatim.
• Underestimating effort: hidden work sinks timelines. Fix: Make AI list dependencies and unknowns; multiply effort by 2× when unknowns exist.
• Scoring drift: rescoring daily creates churn. Fix: Rescore only after a test or new material evidence.
• Fuzzy tests: too many changes at once. Fix: One change, one metric, clear pass/fail threshold.

One-week action plan

1. Day 1: Run the sprint steps 1–3. Leave with a Now/Next/Later list.
2. Day 2: AI-draft the top test brief, edit for your context, schedule the send or page change.
3. Days 3–9: Execute the 7-day micro-test. Daily 5-minute check-in: is the change live? Any blockers?
4. Day 10: Ask AI for a one-page debrief from your metrics. Rescore the list and pick the next test.

Copy-paste prompt: AI Prioritization Copilot

“You are my prioritization copilot. I will paste: (1) a list of short insights and (2) evidence snippets (reviews, emails, notes, metrics). Tasks: 1) Normalize and deduplicate insights; return canonical one-sentence versions grouped by theme. 2) For each insight, score Impact, Effort, and Confidence from 1–5 with one-sentence justifications using only the evidence I provide; if effort has unknowns, apply an Unknowns Multiplier of 2× to Effort. 3) Compute Priority=(Impact*Confidence)/Effort and label each insight Now (top 3), Next (next 5), Later (the rest). 4) Create an Evidence Pack: 2–3 verbatim quotes per insight that support or challenge it. 5) For the top Now item, draft a one-week micro-test brief: exact change, audience, single metric, baseline (state assumptions if needed), pass/fail threshold, risks, and a 7-day checklist. Output two sections: A) CSV with columns: Insight,Theme,Impact,Effort,Confidence,EffortMultiplierApplied,Priority,Label; B) The micro-test brief and checklist in plain language.”

Optional prompt: Effort Map for one idea

“For this single insight [paste], list tasks, skill needed, who can do it, time estimate (min/hrs), dependencies, and unknowns. Flag anything that would double the effort. Return a simple checklist.”

What to expect

• A clean, ranked list you trust — not just gut feel.
• One test you can launch this week with a clear success line.
• Fewer debates; more small wins that stack.

Pragmatic and optimistic: use AI to clear the fog, then move. Score fast, test small, learn weekly. That rhythm compounds.

Quick win (5 minutes)

Grab a notepad. List your top 5 weekly tasks, your monthly budget, and one required integration (calendar or payments). Then paste the prompt below into your AI and get a ranked shortlist, a 90-minute setup checklist, and 30‑day payback math. That’s enough to choose what to trial this week.

Copy‑paste prompt

“Act as my tool‑stack analyst. My weekly tasks are: [list tasks]. Budget: [$/month]. Required integrations: [e.g., Google Calendar, Stripe/PayPal]. My hourly rate (for ROI): [$].

Return:
1) 3 categories (CRM, invoicing/payments, project tracking) with 2–3 practical options per category. For each option, include: monthly cost estimate, setup time (hours), key integrations, learning curve (low/medium/high), one downside, CSV import/export availability, and any native connectors.
2) Score each option using this weighting: Fit to weekly tasks 40%, Integration path incl. CSV/connectors 25%, Time‑to‑first‑value 20%, Total monthly cost 10%, Exit ease 5%. Show a 1–5 score and the weighted score.
3) Recommend the fastest‑to‑value small stack (2–3 tools). Provide a 90‑minute setup checklist and a 7–day mini‑trial script (typical + edge case + data export/import test).
4) Do 30‑day payback math: (minutes saved/week × $/hour × 4) – monthly tool cost. Flag PASS if ≥ $0.
5) List a simple exit plan (how to export and disconnect cleanly).”

Why this works

AI is great at narrowing choices and structuring trade‑offs. You keep control with numbers: time saved, cost, and a clear exit. This turns “shiny features” into “measurable wins.”

What you’ll need

• One‑page task list and a single success goal (save X minutes/week or cut $/month).
• Budget range and 1–2 must‑have integrations (calendar, email, or payments).
• Your hourly rate (or a simple value per hour you’re willing to invest).

Step‑by‑step to a lean stack

1. Get the shortlist: Run the prompt above. Ask follow‑ups until the shortlist references your exact tasks and integrations.
2. Score discipline: Use the weighted grid in the prompt. Anything scoring under 4.0/5.0 is a candidate for “park for later.”
3. 90‑minute setup rule: For the top pick, book 90 minutes. Complete one end‑to‑end workflow (create client → schedule → log time → invoice → receive payment). If you can’t finish, it’s a red flag.
4. Trial script: Run two cases—typical and edge. Edge example: correct an invoice and process a refund. Add a “data exit” check: export a CSV and re‑import it without mangling fields.
5. ROI check: Use this line: minutes saved/week × hourly rate × 4 ≥ monthly cost. If not, move to the next candidate.
6. Decide: If it passes scoring, setup, and ROI, keep it. Otherwise, discard quickly and test the next option.

Insider trick: exit‑first selection

Before you fall in love, verify you can leave. Confirm CSV export/import and a basic connector path on day one. If you can’t get your data out cleanly, don’t go in.

Worked example (numbers you can copy)

• Tasks: onboarding, scheduling, time tracking, invoicing, follow‑ups.
• Budget: $60/month. Hourly rate: $100. Target: save 30 minutes/week.
• AI returns a stack with estimates: save 45 minutes/week; setup 2 hours; cost $48/month.
• 30‑day payback: 45 × $100 × 4 = $18,000? No—convert minutes to hours: 45 minutes = 0.75 hours. 0.75 × $100 × 4 = $300. Payback = $300 – $48 = +$252. PASS.
• Decision: Keep and document a 1‑page SOP for the new flow.

Extra prompts to tighten the process

• Integration smoke test (copy‑paste): “Create a 30‑minute smoke test for my shortlisted stack covering: create contact, schedule via calendar, generate invoice, accept card payment, confirm data sync across apps, export CSV and re‑import. For each step, list expected result, what failure looks like, and how to collect evidence (screenshot or log).”
• Pre‑mortem (copy‑paste): “Run a pre‑mortem on my chosen stack. List 5 likely failure modes (integration breaks, hidden costs, data lock‑in, adoption issues, support gaps) and one mitigation I can execute during the 7‑day trial for each.”
• SOP generator (copy‑paste): “Draft a 1‑page SOP for my final stack covering: who uses it, when, the exact steps, fields to complete, what ‘done’ looks like, and a rollback procedure.”

Common mistakes and quick fixes

• Chasing features → Ask: which feature removes a manual step today? If none, ignore it.
• Free tier lock‑in → Confirm export paths before you import data. Free isn’t free if you can’t leave.
• Over‑integrating → Start with calendar + payments. Add email or automation only if it cuts real steps.
• Long pilots → 7–14 days max with decision gates. No rolling trials.
• No documentation → Write the SOP as soon as the trial passes. It cements habits and reduces errors.

48‑hour action plan

1. Hour 0–1: List tasks, budget, integrations, hourly rate. Run the main prompt. Get a scored shortlist.
2. Hour 2: Pick the top candidate per category. Book 90 minutes per tool.
3. Hour 3–4: Execute the 90‑minute setup rule. Complete one real workflow. Log time, errors, and any workarounds.
4. Hour 5: Run the smoke test and export/import check. Calculate 30‑day payback.
5. Hour 6: If it passes, draft a 1‑page SOP and enable MFA. If it fails, move to the next candidate immediately.

What to expect

• A lean stack of 2–3 core apps that covers 80% of your needs.
• Time savings you can measure within a week.
• Confidence to keep or cut tools based on evidence, not hype.

Closing thought

Use AI to compress research and make numbers visible. Keep trials short, protect your exit, and only adopt what pays back in 30 days. Small, proven wins build a calm, durable workflow.

Try this now (3–5 minutes): Paste the prompt below into your AI. You’ll get a complete, student-ready mini project with a real user, clear milestones, a tight rubric, and a short exemplar you can launch this week.

Copy-paste prompt:

“You are an expert project-based learning designer. Build a 1-page PBL Sprint Kit for learners that tackles [insert a real problem, e.g., reduce cafeteria waste by 25% in 4 weeks]. Include: (1) a driving question and named user, (2) 3 learning outcomes with measurable success criteria, (3) 3 milestones with deliverables and dates, (4) 4 student roles, (5) a concise 10-point rubric (3 criteria x descriptors for Excellent/Acceptable/Insufficient), (6) a 150–200 word exemplar of the final deliverable, (7) a 10-minute stakeholder check-in script and acceptance criteria. Keep language simple and student-facing. Add low-tech options and note adaptations for diverse reading levels.”

Why this works: Real users, clear acceptance criteria, and short sprints make PBL feel like real work and easier to assess. AI gives you the first draft fast; you add the human touch.

What you’ll need:

• A conversational AI (ChatGPT or similar)
• One authentic partner or a realistic client scenario
• A shared folder (Docs/Drive) for brief, rubric, and exemplars
• 10–15 minutes with a stakeholder for feedback

Step-by-step (60–90 minutes prep + 1-week pilot):

1. Frame the job. Write one sentence with a metric and a timeline. Example: “How can our school cut printing costs by 20% this term?” User: school admin.
2. Set outcomes. Pick 3 you can observe: research evidence, prototype/test, public presentation. Add one success criterion to each (e.g., “includes 3 sources and a data table”).
3. Draft with AI. Use the Sprint Kit prompt above. Iterate until the language is plain and student-facing.
4. Calibrate quality (insider trick). Ask AI to create three sample student submissions (Excellent/Acceptable/Insufficient). Use these to norm your scoring and show students the bar.
5. Plan the check-in. Book a 10–15 minute stakeholder slot in Week 1. Their quick rating (1–5) and one improvement note becomes your acceptance test.
6. Equip students. Generate a Milestone 1 checklist and a peer-review form tied to the rubric. Print or share digitally.
7. Launch and loop. Assign roles, run Milestone 1 as a 1‑week sprint, collect submissions, run peer review, score with the rubric, and log one adjustment for next week.

Concrete example (use or adapt):

• Project: School Energy Saver Challenge.
• Driving question: How can our school cut classroom energy use by 15% in 6 weeks without buying new equipment?
• User: Facilities manager.
• Outcomes: (1) Evidence-based causes (audit report), (2) Low-cost prototype (behavioral or scheduling change) with before/after data, (3) 3‑minute stakeholder pitch.
• Roles: Researcher, Data Lead, Prototype Designer, Presenter.
• Milestones: 1) Audit & data snapshot; 2) Prototype & test for 5 days; 3) Final pitch with charts and next steps.
• Acceptance criteria: At least one week of comparative data, a simple cost–benefit note, and a clear ask for scale-up.

Premium shortcuts that save time and raise quality:

• Authenticity triangle: lock three elements up front — real user, real constraint (budget/time), public deliverable (pitch, poster, one-pager).
• Rubric compression: use only 3 criteria (Evidence, Solution Quality, Communication). It speeds scoring and improves feedback.
• Evidence map: require 3 artifacts across the project: a data table, two photos/screenshots, and one stakeholder quote.
• Equity levers: ask AI for two reading levels of the brief and a sentence starter bank for English learners.

More copy-paste prompts (use after the Sprint Kit):

• Calibration set: “Using the rubric above, generate three versions of a student final deliverable on [topic]: one Excellent, one Acceptable, one Insufficient. Keep length 180–220 words and annotate 2–3 sentences explaining why each meets its level.”
• Milestone 1 checklist: “Create a student-facing checklist for Milestone 1 tied to the rubric. Include 8–10 items, simple language, and a self-rating scale (Yes/Almost/Not yet).”
• Comment bank: “Write 30 short teacher feedback comments mapped to rubric levels for Evidence, Solution Quality, and Communication. Max 12 words each.”
• Adaptive scaffolds: “Suggest low-tech, low-cost alternatives for each deliverable and two supports for students who struggle with reading and data (sentence frames, visual templates).”

Common mistakes and quick fixes:

• Scope creep: Too many deliverables. Fix: cap at 3 milestones and 3 rubric criteria.
• Vague outcomes: No metrics. Fix: add one observable measure per outcome (e.g., “3 sources,” “one A/B test,” “two stakeholder comments”).
• Over-automation: AI drafts; you finalize. Keep human scoring and stakeholder input.
• Weak feedback loops: Build a short peer review and a 10‑minute stakeholder check-in into Milestone 1.

1-week action plan:

1. Mon: Generate the Sprint Kit + calibration samples. Post to your folder.
2. Tue: Create the Milestone 1 checklist, peer-review form, and comment bank.
3. Wed: Launch the project, assign roles, start data gathering or research.
4. Thu: Peer review in class; collect quick stakeholder rating and one note.
5. Fri: Score with the rubric, summarize results on one page, adjust next sprint.

What to expect: Clearer student work, faster prep, and tangible artifacts you can show parents and partners within a week. Your second run will be smoother — you’ll refine the rubric and reuse the checklists.

Tell me your age/grade and subject, and I’ll tailor the Sprint Kit prompts and an example project to fit your class.

On your side,

Jeff

Hook

Short answer: yes — AI can map a practical tool stack for your workflow. Quick correction: don’t treat a “must integrate with my bank” as absolute—often a payment processor (Stripe/PayPal) or CSV bank exports are sufficient and much faster to implement. Prioritise integrations by impact, not by label.

Context — why this approach works

AI is best as a research assistant: it narrows choices, highlights trade-offs and saves you hours of browsing vendor pages. You still decide. The goal is quick, measurable wins that reduce friction.

What you’ll need

• A one-page list of weekly tasks and one business goal (save X hours/week or cut Y $/month).
• Budget range per month and 1–2 priority integrations (e.g., calendar + payment processor).
• 30–60 minutes to run the AI session and 7–14 days to test a shortlist.

Step-by-step — how to do it

1. Share your one-page task list and constraints with the AI. Ask for 2–3 options per category (CRM, invoicing/payments, project tracker).
2. Request pros/cons tied to your constraints: cost, setup time, integrations, learning curve, and one downside each.
3. Choose top 1–2 candidates per category and run a short trial (7–14 days). Use a test script: one typical workflow + one edge case.
4. Record metrics: time saved, errors removed, manual steps eliminated, and adoption rate.
5. Keep the tool that meets your goal; otherwise run the next candidate from the shortlist.

Test script (copy-and-paste to use)

Typical workflow: create a new client record, schedule an appointment, log one week of time, send an invoice and process a payment. Edge case: client requests an invoice correction and a refund.

Copy-paste AI prompt (use as-is)

“I am a solo consultant with these weekly tasks: client onboarding, time tracking, invoicing, and client follow-ups. My budget is $60/month. I need Google Calendar integration and ability to accept card payments (Stripe or PayPal OK). Recommend 3 categories (CRM, invoicing/payments, project tracker) and list 2–3 practical options per category. For each option, give: monthly cost estimate, setup time (hours), key integrations, likely learning curve (low/medium/high), one major downside, and whether quick CSV import/export or connector exists. Then recommend one small stack for fastest implementation and explain why.”

Common mistakes & fixes

• Mistake: Chasing every feature. Fix: Prioritise features that remove manual steps.
• Mistake: Skipping integration checks. Fix: Test one real transaction and one calendar event during trial.
• Mistake: Long pilots. Fix: Use focused 7–14 day trials with clear metrics.

Worked example (quick)

Solo consultant wants low-cost stack. AI suggests: simple CRM (notes+follow-ups), invoicing that uses Stripe, and checklist-based project tracker. Trial shows one app saved 45 minutes/week and removed duplicate data entry — winner.

7-day action plan

1. Day 1: Create one-page task list and goal.
2. Day 2: Set budget and primary integrations.
3. Day 3: Run the AI prompt above; get shortlist.
4. Day 4: Sign up for top candidates (fast setup only).
5. Day 5–11: Run test script, capture metrics.
6. Day 12: Decide, keep or iterate.

Closing reminder

Use AI to shrink choices, not to swap judgement. Small, tested steps win — aim for measurable improvements and one clear decision at the end of each trial.