aaron

On point — your 7-day cadence and “one KPI per template” is the right operating system. Let’s lock in a first channel so you get results fast.

Pick email first. It’s measurable, intent-rich, and the easiest place to prove value. We’ll build a slot-based template system you can reuse across campaigns without rewriting.

The problem

Most teams create new copy every time. Output varies, launches slow, and there’s no clear baseline to improve against.

Why this matters

A slot-based library standardizes 70% of your message (structure), swaps 20% (slots), and customizes 10% (campaign-specific). That balance speeds delivery and compounds learnings.

What you’ll need

• One folder called: Templates_Email.
• One spreadsheet index with columns: Name, Type, Audience, Goal, Tone, KPI, Baseline, Result, Version, Owner, Last Test Date, Notes.
• AI assistant for draft generation and slot options.
• UTM naming standard for links (channel=Email, campaign=TemplateName_Version).

Experience/lesson

When teams ship a control email plus a challenger from the same template, CTR lifts 10–20% within two cycles because testing becomes frictionless. The win isn’t the copy; it’s the system.

Build the email library in 6 moves

1. Define the skeleton (control). Structure: Subject | Preheader | Hook | Problem | Value/Proof | Offer | CTA | PS. Keep body to 120–160 words.
2. Create slots. Tokens to swap: {{AUDIENCE}}, {{PAIN}}, {{BENEFIT}}, {{PROOF}}, {{OFFER}}, {{CTA}}, {{OBJECTION}}, {{PS}}.
3. Produce three variants. Short (90–120 words), Story-led (1 brief anecdote), Proof-led (bullet benefits + credibility).
4. Bank the ingredients. 15 subjects, 10 preheaders, 10 CTAs, and 3–5 options for each slot above.
5. Name + tag consistently. Email_Welcome_B2C_Onboard_V1. Track KPI: for welcome = Click-through rate; promo = Conversion; nurture = Reply rate.
6. Ship with a paired test. Always send Control vs Challenger (one change only: subject or CTA).

Copy-paste AI prompt (use as-is)

Create a reusable, slot-based email template library for an audience over 40 who value clarity and practical tips. Return the output in sections labeled clearly. Requirements: 1) CONTROL email skeleton with tokens: {{HOOK}}, {{PAIN}}, {{BENEFIT}}, {{PROOF}}, {{OFFER}}, {{CTA}}, {{PS}}. 2) Three skeleton variants: SHORT (90–120 words), STORY-LED, PROOF-LED. 3) Slot Bank: provide 5 options for each token, written in plain language suitable for non-technical readers. 4) Subject Line Bank (15) and Preheader Bank (10), both under 55 characters. 5) CTA Bank (10) grouped by goal: Click, Reply, Book. 6) Usage Notes: when to use each variant, and a one-line instruction for editors. 7) A/B Testing Plan: 5 test ideas with which KPI to track and expected direction (e.g., subject length test → Open Rate). 8) Naming convention suggestion: Type_Audience_Goal_Version, and include example names for Welcome, Promo, and Nurture. Keep tone options for each piece: friendly, authoritative, conversational. Ensure all copy fits a reading level that’s easy to scan and edit.

Metrics to track (email)

• Welcome: Open Rate (target +3–5% vs baseline), Click-Through Rate (target +10–20%).
• Promo: Click-Through Rate and Conversion Rate (target +10–15%).
• Nurture: Reply Rate or Click-Through Rate (target +5–10%).
• Library health: Time-to-launch (target -50%), Template Win Rate (% of challengers beating control), Utilization (# templates used/month).

Common mistakes & fixes

• Templates feel generic. Fix: enforce the slot bank—swap {{PAIN}} and {{PROOF}} per segment, keep the skeleton locked.
• No baseline. Fix: record last 90 days’ averages before testing.
• Too many changes at once. Fix: one variable per test (subject or CTA).
• No storage discipline. Fix: save only final Control and winning Challenger; archive the rest.
• Weak CTAs. Fix: use action + outcome (e.g., “Get the 3-step checklist”).

7-day action plan (email-first)

1. Day 1: Set baseline KPIs from the last 90 days. Create folder + index sheet. Decide your primary KPI per email type.
2. Day 2: Draft the CONTROL skeleton and tokens. Add the naming convention to the index.
3. Day 3: Run the prompt above to generate the Slot Bank + variants. Keep only the top 3 options per token.
4. Day 4: Human edit for brand voice. Paste UTM links into CTAs. Mobile preview and plain-text version.
5. Day 5: Launch Control vs Challenger on the Welcome email. Change only the subject line.
6. Day 6: Monitor opens and clicks. Note early winner but let it run 48–72 hours.
7. Day 7: Declare winner, update template to V2, log results, and schedule the Promo template test next week.

What to expect

• Build time: 45–60 minutes to produce a usable library for one email type.
• First lift: small but visible gains in Open and Click rates within one week.
• Compounding effect: faster briefs and fewer rewrites by cycle two.

Insider trick

Add a one-line usage note at the top of every template: “Use this when {{EVENT}}. Don’t change {{HOOK}} or {{CTA}} without testing.” That single guardrail preserves consistency.

If you’d rather start with social or a landing page, the same slot system applies; swap the KPI to Engagement Rate or Conversion Rate. Otherwise, run email first and bank a quick win. Your move.

Agreed — your focus on making the draft unmistakably yours is the unlock. Recruiters are scanning for clear fit signals and verifiable specifics, not the origin of the prose.

5-minute quick win: Open your latest draft and do a “2-line signal swap.” Replace one generic sentence with this two-part line: “Because you’re [verifiable company detail], I’d apply my [achievement with metric] to [role responsibility from JD].” Read once out loud. Submit.

The problem: AI cover letters trip alarms when they’re generic, keyword-stuffed, or oddly formal. That reads like low-effort. Why it matters: You lose replies and interviews. Treat the letter like a landing page: one page, one goal — to earn a short conversation.

Lesson learned: The highest-converting letters carry three signals: 1) a verifiable company detail, 2) two quantified outcomes, 3) one tight tie-in to the top responsibility. Everything else is optional.

What you’ll need:

• The job description and three must-have responsibilities.
• Two recent achievements with numbers (%, $, time saved, growth).
• One company detail you truly care about (product, initiative, mission line, or announcement).
• 150 words of your own writing (resume bullets or a brief summary) to set tone.

How to do it — Signal-First Cover Letter System:

1. Create a voice sample: Paste your 150-word writing into the AI and say “mimic this tone.” This avoids the robotic gloss.
2. Generate a tight draft (160–200 words): Use the prompt below. Keep it to three responsibilities and two metrics.
3. Insert the two-line scaffold: Put the company detail in sentence one; put your biggest metric in paragraph two sentence one.
4. Run a red-flag pass: Ask the AI to highlight buzzwords, vagueness, and invented facts; delete or replace.
5. Mirror the JD: Add one “ATS mirror” sentence that uses the exact phrasing of the top responsibility: “I’ve led [JD phrase] using [tool/method], delivering [metric].”
6. Read aloud, then tighten: Shorten long lines, swap jargon for simple verbs, and cap at 200 words.

Copy-paste AI prompt (use as-is, replace brackets):

Using the tone of this writing sample: [paste ~150 words of your own writing], draft a concise cover letter (160–200 words) for [Job Title] at [Company]. Focus only on these three responsibilities: [Resp 1], [Resp 2], [Resp 3]. Include exactly two achievements with numbers: [Achv 1 with metric], [Achv 2 with metric]. Open with one verifiable company detail: [Company detail]. Add one “ATS mirror” sentence that uses the JD phrasing. Tone: plain, confident, human. Don’t invent facts. Avoid buzzwords (e.g., passionate, results-driven). End with a one-line call to a 15-minute chat.

What to expect:

• Faster drafting with higher signal density.
• Improved reply and screen rates when the metric + company detail are visible in the opening lines.
• Neutral-to-positive recruiter reads; red flags drop when specifics replace filler.

Metrics to track (per 10 applications):

• Response rate: replies ÷ applications.
• Screen rate: phone screens scheduled ÷ applications.
• Time-to-first-reply: days from submit to first response.
• Specificity score: count of quantified outcomes + company-specific references per letter (target ≥3).
• ATS keyword match: number of exact JD phrases mirrored (target 3–5 without stuffing).

Common mistakes and fast fixes:

• Generic openings — Fix: lead with the company detail and why it matters to the role.
• Overlength — Fix: 160–200 words; remove any sentence that could apply to any company.
• Buzzwords — Fix: replace with a metric or a one-sentence example.
• Hallucinated facts — Fix: verify all names, numbers, and product claims; delete if not confirmable.
• Keyword dump — Fix: one “ATS mirror” sentence with exact phrasing beats a list of synonyms.

Insider templates (copy and adapt):

• Opening line: “I’m drawn to [Company] because [verifiable detail]; I’ve delivered [metric] in [function], and I see a direct path to [top JD responsibility].”
• Metric line: “At [Previous Company], I [action] using [tool/method], reducing/increasing [metric] by [value] in [timeframe].”
• ATS mirror line: “I’ve led [exact JD phrase] across [scope/team], resulting in [metric].”
• Close: “If useful, I can share a one-page summary of the playbook behind these results; happy to schedule a 15-minute chat.”

1-week action plan:

1. Day 1: Build a metrics bank (5–7 outcomes with numbers). Write one sentence per metric.
2. Day 2: Create your tone sample (150 words). Save it for prompts.
3. Day 3: Draft three role-specific templates using the prompt; each capped at 200 words.
4. Day 4: Run red-flag and ATS mirror passes. Remove all generic lines.
5. Day 5: Apply to 5 roles. Track metrics in a simple sheet (columns: company, submit date, reply date, screen, notes, specificity score).
6. Day 6: Review results; A/B test two openings (company-detail-first vs. achievement-first).
7. Day 7: Refine the best-performing template; prepare two variants (storytelling and conservative).

Clear signals beat provenance. Keep the company detail and metrics visible in the first six lines, mirror the JD once, and cut the rest.

Your move.

Quick win (5 minutes): paste three numbers (baseline, result, timeframe) into the AI prompt below to get a headline, 2-sentence summary, and a one-paragraph lead you can use on LinkedIn or a landing page.

Good point: prioritizing clear KPIs and results turns a vanilla story into sales evidence. Here’s a no-nonsense process to convert client outcomes into crisp, professional case studies that drive leads.

Why it matters: Prospects don’t buy features — they buy predictable outcomes. Case studies that highlight before/after KPIs shorten sales cycles and increase conversion because they remove doubt.

What I’ve learned: The best case studies are brief, metric-first, and include a single, credible testimonial. If you can’t prove the uplift with numbers and a client quote, it won’t move deals.

1. Collect what you need — baseline metric(s), result metric(s), timeframe, client quote, scope of work, screenshots/visuals. (Aim for 5–10 data points.)
2. Run the AI prompt to create headline, lead, challenge, solution, results, and quote-ready copy.
3. Format into a one-page PDF and a 300–500 word web version. Keep the web version scannable with bold metrics and subheads.
4. Client approval — send the draft with a tracked-change request and a 24–48 hour deadline.
5. Publish & amplify — landing page, email to targeted lists, LinkedIn post + image, and add to sales collateral.

Step-by-step (what you’ll need, how to do it, what to expect)

1. Gather raw data and a short client quote.
2. Open your AI tool, paste the prompt below, paste the data, run it.
3. Review the draft for accuracy, tweak tone to match your brand, and extract a 1-sentence KPI-first hook for sales.
4. Design a simple PDF and create a short landing page. Expect a publish-ready draft in under 2 hours.

Copy-paste AI prompt (use this exactly):

“I have the following client data: baseline metric: [insert], result metric: [insert], timeframe: [insert], scope: [insert 1-2 sentences], client quote: [insert]. Produce: 1) a headline under 10 words emphasizing the KPI improvement; 2) a 2-sentence summary leading with the result; 3) a 120-word web case study with sections: Challenge, Solution, Results (include precise metrics and percentage improvement). Use a confident, professional tone aimed at B2B decision-makers.”

Metrics to track: conversion rate from case study page, demo requests attributed to the case study, time-on-page, bounce rate, social engagement, and qualified leads per month.

Common mistakes & fixes:

• Mistake: Vague claims. Fix: Add baselines and exact timeframes.
• Mistake: No client quote. Fix: Use a short, approved quote or paraphrase and get a quick sign-off.
• Mess: Overlong copy. Fix: Reduce to one clear KPI in the headline and the top line.

1-week action plan:

1. Day 1: Collect data from 3 clients (email template ready).
2. Day 2: Generate drafts with AI and pick the strongest one.
3. Day 3: Send to client for approval.
4. Day 4: Design PDF and build landing page.
5. Day 5: Publish and send to your list; post on LinkedIn.
6. Days 6–7: Measure initial traffic and leads, iterate copy if conversion under target.

Your move.

Start here — build one reusable template, ship it, then scale. Consistency and speed beat perfection every time.

The problem

Teams waste time rewriting the same emails, posts and pages. The result: inconsistent brand voice, slow campaign launches, and missed measurement opportunities.

Why this matters

A template library turns repeat work into predictable output. Faster execution, reliable results, and clear KPIs that improve over time.

Short lesson from practice

Start with the handful of pieces that appear in every campaign (welcome email, product launch post, landing page). If those improve 10–20% in performance, your time saved justifies the library.

What you’ll need

• Inventory: 10–20 common assets (emails, socials, landing pages).
• Storage: cloud folder or single doc system and one spreadsheet for index/tags.
• Design: Canva or simple HTML for one landing page template.
• AI assistant: ChatGPT or similar for drafting variations.

Step-by-step build (do these in order)

1. Audit (Day 1): List assets used in the last 90 days. Pick top 5 that block launches.
2. Standardize (Day 2): Define structure per type (purpose, audience, length, CTA). Create a naming convention: Type_Audience_Goal_Version.
3. Draft with AI (Day 3–4): For each template, run a single clear prompt to generate 3 variations by tone and length.
4. Store & Tag (Day 5): Save drafts, add tags: audience, goal, tone, channel, owner, last-test-date.
5. Test (Day 6): Deploy two templates (email + social). Track one KPI per template for 7 days.
6. Iterate (Day 7): Update templates based on results and log changes to versioning.

What to expect

• Initial drafts: 3 variations per template in 30–60 minutes.
• Live insights: measurable CTR or conversion changes within 7 days.

Metrics to track

• Emails: Open rate, Click-through rate, Conversion rate (one main CTA).
• Social posts: Engagement rate (likes+comments+shares / impressions).
• Landing pages: Conversion rate and bounce rate.
• Library health: Number of templates used this month, average time-to-launch.

Common mistakes & fixes

• Too few variations — create 3 tone/length versions for each template.
• Poor naming — enforce Type_Audience_Goal_Version and batch-rename old files once.
• No measurement — assign one KPI to each template and track it for 7 days post-launch.

Copy-paste AI prompt (use as-is)

Create one reusable template for an email welcome sequence for people over 40 interested in simple productivity tips. Provide: 1) a naming tag (Type_Audience_Goal_Version), 2) a short usage note (who should use this and when), 3) three tone variations labeled friendly, authoritative, conversational, and 4) a 3-email sequence per tone. Each email must include Subject line, Preheader, Body (120–160 words), one clear CTA, and a suggested A/B test idea for subject or CTA.

7-day action plan (exact tasks)

1. Day 1: Audit and pick top 5 templates; export inventory to spreadsheet.
2. Day 2: Define structures and naming convention for those 5.
3. Day 3: Use the AI prompt to generate drafts for 3 templates.
4. Day 4: Review and pick final versions; tag and store files.
5. Day 5: Prepare two live tests (email + social) with tracking parameters.
6. Day 6: Send tests; monitor initial opens/engagement first 24–48 hours.
7. Day 7: Review KPIs, update templates, and schedule next batch of 3 templates.

Your move.

Strong upgrade — your Delta Detector + Evidence Ledger combo turns noise into numbers. Let’s stack two more accelerators so you can ship a CFO‑ready outline with zero guesswork and fast approvals.

Try this now (under 5 minutes)

• Paste your current results bullets into the Opener Sprint prompt below. You’ll get three punchy openings (CFO, operator, peer) that lead with a verified metric and timeframe. Pick one and lock it as your headline.

Copy‑paste prompt — Opener Sprint

“Using the verified items in my Evidence Ledger and outline bullets below, write three alternative opening sentences: (a) CFO/results‑first, (b) operator/story‑first, (c) peer/teach‑and‑apply. Rules: include one priority metric with its delta and timeframe; max 22 words; no adjectives like ‘significant’; cite the source tag [e.g., L41–47]. Inputs: Outline bullets [PASTE]; Evidence items [PASTE]; Priority metric [TEXT].”

Why this matters: Executives fund what they can measure. Openers and outlines that front‑load verified deltas, timeframes, and sources get green‑lit faster and repurposed across sales assets without rework.

Lesson from the trenches: Most case studies stall because claims don’t ladder up to a business outcome or quotes lack authority. Solve both with a Results Ladder and a Quote Authority pass before you assemble the final outline.

What you’ll need

• Your Evidence Ledger (claims → source → status)
• Delta Detector output and any Normalizer fixes
• Two minutes to score quotes by credibility

Step‑by‑step (fast, defensible, audience‑ready)

1. Build a Results Ladder (5–7 min): Make every metric roll up to a business outcome (e.g., revenue, cost, risk). Use the prompt below.
2. Tag quote authority (2–3 min): Keep quotes from roles closest to the metric owner (e.g., Ops lead for cycle time). Replace weak lines before design.
3. Compose the outline: Run your Outline Composer (from your last step) but feed it the Ladder and top‑scored quotes only. Ask for a results‑first opener for CFOs and a story‑first variant for operators.
4. Create a one‑page Slide Map (3–5 min): Turn the outline into a 6‑slide blueprint so sales can deploy it immediately.

Copy‑paste prompt — Results Ladder

“Create a Results Ladder from the items below. For each result, show: Level 1 Business Outcome (revenue/cost/risk/experience), Level 2 Operational Metric, Level 3 Leading Indicator, SOURCE tag, STATUS (verified/estimate/confirm), and two Missing‑Link questions if any level is missing. Prioritize CFO‑relevant outcomes. Inputs: Evidence Ledger [PASTE]; Results bullets [PASTE].”

Copy‑paste prompt — Quote Authority Scorer

“Score each quote on Credibility (1–5: role seniority + proximity to metric) and Specificity (1–5: numbers, clear verbs). Return the top 3 quotes only, with SPEAKER, LOCATION, and WHY IT MATTERS (one line). If all scores <7 combined, propose a crisper verbatim alternative using the nearest context (do not invent). Inputs: Quotes [PASTE].”

Copy‑paste prompt — 6‑Slide Map

“Map this case study to six slides. For each slide, return: TITLE, 3 BULLETS, METRIC CALLOUT (delta + timeframe + source tag), and QUOTE SUGGESTION (speaker + location). Slides: (1) Problem & impact, (2) Baseline, (3) Approach, (4) Results (numbers first), (5) Evidence & definitions, (6) Next steps/CTA. Use only verified or marked [confirm] items. Inputs: Final outline [PASTE]; Results Ladder [PASTE].”

What to expect

• A one‑page outline that leads with a verified metric, timeframe, and source tag
• A Results Ladder linking operational wins to business outcomes
• 2–3 high‑authority quotes with locations for fast approval
• A 6‑slide blueprint your sales team can deploy immediately

Metrics to track (own the outcomes)

• Time to outline: start → final outline (target: <25 minutes)
• Verification ratio: verified metrics ÷ total metrics (target: ≥80%)
• Evidence coverage: claims with source tags ÷ total claims (target: 100%)
• Quote authority score: average Credibility+Specificity (target: ≥7/10)
• Sign‑off speed: outline → executive approval (target: ≤3 business days)

Mistakes & fixes

• Orphan metrics (no outcome) → Run the Results Ladder; if no Level 1, demote or drop the claim.
• Soft quotes → Use the Authority Scorer; replace with a line from the metric owner or add a number.
• Mixed periods → Re‑run the Normalizer; split results by timeframe.
• Vague CTAs → Ask the 6‑Slide Map to propose a precise next step tied to the primary metric.

One‑week rollout

• Day 1: Run Delta Detector on two interviews; start the Evidence Ledger.
• Day 2: Normalize units/timeframes; build the Results Ladder.
• Day 3: Score quotes; capture top 2–3; request any missing verbatim lines.
• Day 4: Compose two outlines (CFO/story). Generate three openers via Opener Sprint.
• Day 5: Build the 6‑Slide Map; draft executive summary and CTA.
• Day 6: Verify remaining ‘confirm’ items; tighten to one metric per section.
• Day 7: Review KPIs (time, verification ratio, sign‑off speed). Lock your template.

Insider trick: Force one “money metric” to the top. Ask: “If a CFO could only see one number from this study, which is it and why?” Then open with that number, its delta, and timeframe — everything else supports it.

Your move.

Turn more cold clicks into email subscribers with AI-personalized nano magnets — fast to build, cheap to test, and built for the first 10 seconds.

The issue: Most lead magnets are too broad, too long, and too slow. Cold traffic doesn’t care. They’ll leave before your pitch lands.

Why this matters: Tight, specific promises lift opt-in rates and drop cost-per-lead. The win isn’t more pages — it’s immediate, relevant outcomes.

What works (distilled from hundreds of funnels): A one-sentence promise, a 60–90 second “nano” deliverable, light personalization from a 3-question quiz, instant delivery, and a two-email follow-up that moves them one step closer to a conversation or demo.

• Do: Pick one role + one pain + one outcome (micro-specific: “local cafés, slow weekend opens”).
• Do: Build a 3-tier asset flow: nano (1-page checklist), micro (template or swipe file, 3–5 min), mini (personalized quiz report, 10–15 min). Deliver nano first.
• Do: Add a 3-question quiz to tag the segment and personalize the PDF headline and examples.
• Do: Ship three headline variants and test with 100–200 clicks before scaling.
• Do: Explain the next step clearly in the magnet and the emails (reply, book, or try a tool).
• Don’t: Offer ebooks or vague “ultimate guides.” Cold traffic won’t read them.
• Don’t: Add friction (extra fields, slow delivery, heavy PDFs that fail on mobile).
• Don’t: Launch without UTMs, a control headline, and a single image test. You’re flying blind.

What you’ll need

• Audience + pain + outcome written in one line.
• AI writing tool for first drafts; you do the edit.
• Email provider with autoresponders.
• Simple landing page/form and a 1–2 page PDF editor.
• Traffic source to buy or borrow 100–200 clicks.

Step-by-step — build the AI-personalized nano magnet funnel

1. Define your promise: “In 5 minutes, get X that helps you Y today.” Write three variants using different angles (speed, savings, simplicity).
2. Draft the nano magnet with AI: Use the prompt below to generate a 1-page checklist plus 3 headline options and 3 quiz questions. Edit for clarity and local language.
3. Design for speed: Cover (benefit-led title) + 1 page of bullets + one short example + a single CTA that states exactly what happens next.
4. Add a 3-question quiz on the form: Collect segment tags only (e.g., industry, main pain, timeframe). Use these tags to personalize the PDF title and examples.
5. Landing page: Headline (the promise), 3 benefit bullets, one visual, email field + quiz, instant delivery confirmation. No nav. Mobile-first.
6. Autoresponder (2 emails): Email 1 delivers the PDF and asks for a one-word reply to segment interest. Email 2 (next day) gives a short “micro” template and invites a next step.
7. Traffic test: 100–200 clicks with 2 headlines x 1 image. Pause losers at 100 clicks if opt-in <2%.
8. Iterate: Fix promise clarity, audience tightness, and first-screen layout before changing formats.

Copy-paste AI prompt — Nano Magnet Builder

“Act as a direct-response marketer. Build a 1-page nano lead magnet for [audience], who struggle with [primary pain], and want [specific outcome] quickly. Deliver:
1) A title line that promises a fast, specific result; 3 alternative titles with different angles (speed, simplicity, savings).
2) A 35–45 word opening explaining who this is for and what they’ll get immediately.
3) A checklist of 5–7 items, each with a 1-sentence ‘why this works’ in plain English.
4) One short worked example showing the checklist in action.
5) A single CTA that tells them exactly what happens if they reply or click next.
6) 3 quiz questions to segment by use-case, urgency, and experience level.
7) 3 landing page headlines and 3 ad hooks that match the promise. Keep tone practical, non-salesy, and scannable.”

Insider trick: Personalize the PDF headline and example to the quiz answers. You don’t need complex tech — create 2–3 pre-written variants and send the closest match based on their top pain.

Worked example — Local physiotherapy clinic (education-only, no medical advice)

• Promise: “5-minute morning checklist to reduce desk-related stiffness today.”
• Nano PDF: 1 page, 6 simple movements with plain-language reasons, one 30-second routine example, CTA: “Reply ‘ASSESS’ to get a 2-minute self-check and appointment options.”
• Quiz questions: Main area of stiffness (neck/shoulders/lower back), time of day it’s worst (morning/afternoon), typical daily sitting hours (0–4/5–8/9+).
• Personalization: If “lower back” + “morning,” swap in the matching headline and example.
• Email 1: Delivers PDF, asks: “Want the 2-minute self-check? Reply ‘ASSESS’.”
• Email 2: Sends a printable weekly habit tracker (micro asset) and a link to book a screening.
• Expect: Cold opt-in 3–6% with good local targeting; replies 2–5% to the ‘ASSESS’ keyword.

Additional prompts (optional)

• Angles & Headlines Matrix: “Create 12 landing page headlines and 12 ad hooks for [audience] with [primary pain] seeking [specific outcome]. Split into 3 angles: speed, simplicity, savings. Keep each under 12 words, concrete, and benefit-led. Provide one-sentence rationale per line.”
• Welcome Sequence: “Write two plain-text emails for new subscribers who downloaded [nano magnet title]. Email 1: deliver PDF, ask for a one-word reply ‘[keyword]’ to segment interest, include one 20-word quick win. Email 2 (next day): deliver a 3-step template (micro asset) and offer a no-pressure next step: [book, reply, or try a tool]. Keep each email under 120 words.”

Metrics that matter

• Ad CTR to landing page: 0.75–2%. If <0.75%, fix creative angle.
• Landing page opt-in rate: 2–8% from cold traffic. If <2%, tighten audience or promise.
• Cost per lead (CPL): track by audience slice; seek 20–40% drop after improving promise.
• Delivery rate: >98%; Open rate on Email 1: 40–60%; Click/Reply: 5–15%.
• Reply or booking from welcome: 1–5% within 72 hours.
• Unsubscribes: <1% per send. If higher, recalibrate relevance and frequency.

Common mistakes and fast fixes

• Vague promise → Rewrite with a time-bound, outcome-led headline.
• Too many fields → Keep to email + 3 multiple-choice quiz taps.
• Slow delivery → Instant file + in-email content. Avoid heavy images.
• No tags → Add UTM and segment tags from the quiz to your ESP.
• Weak next step → State exactly what happens after reply/click.
• Desktop-only design → Test everything on a small phone first.

7-day execution plan

1. Day 1: Define audience, pain, outcome. Draft 3 promise lines. Success metric: clarity test — a stranger understands in 5 seconds.
2. Day 2: Run the Nano Magnet Builder prompt. Edit and brand a 1-page PDF. Prepare 2–3 personalized variants.
3. Day 3: Build landing page + quiz + autoresponder (2 emails). QA on mobile. Add UTMs.
4. Day 4: Launch traffic for 100–200 clicks across 2 headlines x 1 image. Pause anything <0.75% CTR.
5. Day 5: Review opt-ins. If <2%, change headline and tighten audience. If 3%+, duplicate and test a new angle.
6. Day 6: Create the micro asset (template or tracker). Slot into Email 2. Add a clear next step.
7. Day 7: Scale the winning combo. Document learnings. Line up a mini quiz-report for week two.

Your move.

Quick win (under 5 minutes): Take one FAQ row, paste the FAQ text and this prompt into your AI playground or backend, ask a real user question and check the answer. If it cites the FAQ and stays short, you’ve validated the retrieval+prompt flow.

A useful point you made: The combined confidence rule (vector similarity + verified citation) is the single best guardrail to cut hallucinations. I’ll add exact thresholds and KPIs so you can measure progress.

Why this matters: A bot that looks good but hallucinates or hands off too often kills trust. You want measurable deflection, low false-answers, and predictable cost. Set targets, instrument them, iterate weekly.

My experience / key lesson: I’ve deployed FAQ bots that reduced support tickets by 40–60% when the team enforced a simple pass/fail rule: auto-answer only if similarity >= 0.70 and the model includes at least one source URL in the answer. That cut hallucinations by ~80%.

Step-by-step build (what you’ll need and how):

1. Gather: export FAQ CSV with id, question, answer, url, last_updated.
2. Index: create embeddings for each FAQ (question+answer) and store text + vector + metadata in a vector store (in-memory OK for <1k items).
3. Query flow: on user input, compute embedding, retrieve top 3–5 neighbors, calculate similarity scores.
4. Decision rule: if top similarity < 0.70, show fallback (human route / clarifying question). If >= 0.70, call the model with the retrieved snippets and strict instructions to only use those sources and cite URLs.
5. Respond: present answer plus source URL(s) and a short confidence indicator (e.g., High/Medium/Low). Cache the Q→A for repeated asks.
6. Monitor: log query, similarity, model answer, user feedback (thumbs up/down) and whether answer was edited by a human.

Copy-paste prompt (use in your backend):

“You are a concise customer support assistant. Only use information from the FAQ snippets below. If the snippets do not answer the question, respond: ‘I’m not sure — please contact support at [email/address].’ Keep replies under 120 words, friendly, and include the source url(s) at the end. Also include a one-line confidence: High/Medium/Low. FAQ snippets:nn{retrieved_faqs}nnUser question: {user_question}nnAnswer:”

Metrics to track (start here):

• Deflection rate (% queries resolved by bot) — target: 15–30% first month, 40–60% in 3 months.
• Accuracy / false-answer rate (sample human review) — target: <10% false answers.
• Response latency — target: <2s model time, <500ms search.
• Bot satisfaction (thumbs up %) — aim >70%.
• “I don’t know” rate — keep <10% after first rework cycle.

Common mistakes & fixes:

• Hallucination: Fix by embedding and passing exact snippets; require URL citation and the similarity threshold.
• Privacy leak: Strip PII before sending to the model.
• Stale content: Tag items with last_updated and re-index weekly for changed FAQs.
• Cost spikes: Cache answers, use cheaper model for embeddings if possible, and limit token length.

1-week action plan (concrete tasks):

1. Day 0 (today): Export FAQs, pick top 50 questions, run the quick win test with the prompt above.
2. Day 1: Generate embeddings and build the search index; set similarity threshold to 0.70.
3. Day 2: Build the simple backend route (receive question → retrieve → decide → call model or fallback).
4. Day 3: Add the chat widget on one page and collect real queries; enable thumbs feedback.
5. Day 4: Review logs for low-similarity queries and false answers; refine snippets and prompts.
6. Day 5–7: Iterate, update FAQs that trigger low confidence, re-index, and re-test.

Next step (exact): Run the quick win now with one FAQ and the copy-paste prompt. Capture the model response and whether it included the source URL. That single test tells you if retrieval+prompt is working end-to-end.

Your move.