aaron

Quick win: copy your current sheet (even if it’s rough) and run the prompt below to auto-score every relationship and surface the top 5 targets to verify next. You’ll go from an unranked list to a prioritized hit list in under 5 minutes.

The problem — a flat list of names hides what matters: strength of relationship, recency, reciprocity, and who sits at the center of the ecosystem. That’s why maps look busy but don’t drive action.

Why it matters — partners and channels can compress sales cycles and open entire segments, but only if you focus on high-signal, high-centrality nodes. Weight the signals, or you’ll chase press noise.

Lesson from the field — treating partnerships like pipeline works: score the signals, find hubs, and trigger targeted outreach. Expect 30–50% of AI-suggested links to be weak; your edge is how fast you separate noise from moves you can bank.

What you’ll need

• Your spreadsheet with columns: Company, RelatedOrg, RelationshipType, EvidenceNote, EvidenceType, EvidenceDate.
• An AI chat tool and 30–60 minutes for the first pass; 10–15 minutes weekly to maintain.
• Verification sources you can access: news search, company sites, partner directories, job postings, product docs.

Copy-paste prompt — score and shortlist

“You are my ecosystem analyst. Input is CSV with columns Company, RelatedOrg, RelationshipType, EvidenceNote, EvidenceType, EvidenceDate (YYYY-MM-DD). For each row, apply this scoring rubric: +3 official partnership announcement; +2 product integration docs or partner directory listing; +2 marketplace/co-sell listing; +1 investor overlap; +1 shared customer case study; +1 executive quote from either company; +1 multiple independent sources (>=2 distinct types); -2 rumor/speculative language; -1 if the latest evidence is older than 18 months; -1 if evidence is only a single PR pickup with no other signals. Add fields: SignalCount, RecencyDays, Reciprocity (Yes/No if both companies mention each other), Score (0–10), Confidence (High >=7, Medium 4–6, Low <=3), NextAction (Verify, Outreach, Monitor), and a one-sentence Rationale. Return results sorted by Score within each Company, show only the top 10 per Company.”

Step-by-step

1. Normalize names (5–10 min): ensure each organization is consistent (e.g., “AWS” → “Amazon Web Services”). If needed, ask AI: “Unify these organization names into canonical forms and list common aliases; return CanonicalName, Aliases, Confidence.” Update your sheet.
2. Score and triage (5–10 min): run the scoring prompt on your CSV. Flag High (>=7) for immediate action, Medium to verify, Low to monitor.
3. Verify top hits (10–20 min): for each High, confirm two different signals (e.g., partner directory + press release). Update EvidenceNote, EvidenceType, and EvidenceDate. Downgrade anything that fails the two-signal rule.
4. Find hubs (5–10 min): compute simple centrality: count how many times each RelatedOrg appears across your companies. High count = hub. Prioritize hubs that also have High confidence. You can ask AI: “From these edges (Company–RelatedOrg), return top hubs by degree and flag any that connect 3+ of my seed companies.”
5. Decide the play (5–10 min): for each High-confidence hub, pick one: Co-sell (if marketplace/partner listing present), Integration (if API/docs present), Warm intro via investor overlap, or Competitive watch (if it’s a competitor hub).
6. Draft outreach (5–10 min): use AI to create three concrete angles based on your signals. Prompt: “For [TargetOrg], craft 3 concise outreach angles referencing [Signals] and [Shared Customers/Investors]. Include subject lines and a 2-sentence opener.” Paste your best into your CRM or email.
7. Set a refresh loop (5 min): add a Last Verified date and set a weekly 10–15 minute window to update High/Medium rows and re-run scoring.

What to expect

• A ranked, defensible map showing which relationships are real and recent — not just plausible.
• 3–5 outreach-ready targets within a week, plus a shortlist of hubs worth deeper alignment.
• Faster decisions: who to partner with, who to monitor, and where to allocate BD time.

Advanced prompt — entity resolution at scale

“Resolve and deduplicate these organization names. Output CanonicalName, Aliases, ParentCompany (if applicable), and Confidence. Treat variants (e.g., Google Cloud vs. GCP) as one entity. Flag subsidiaries separately if they operate distinct partner programs.”

Metrics to track (weekly dashboard)

• % High-confidence relationships (High / total) — target 30–50% after verification.
• Median RecencyDays for High — keep under 180 days.
• Hub concentration — % of edges accounted for by top 5 orgs; rising concentration indicates where leverage sits.
• Verification cycle time — median minutes from suggestion to verified/downgraded.
• Outreach yield — meetings booked / High-confidence targets attempted.
• False-positive rate — % downgraded after verification; push this under 25% over time.

Common mistakes and quick fixes

• Overweighting press. Fix: require a second, different signal (docs, directory, marketplace, job post) before High.
• Ignoring name variants. Fix: run entity resolution first; it boosts hit rates and reduces duplicates.
• No reciprocity check. Fix: prioritize when both companies acknowledge the relationship.
• Chasing large hubs only. Fix: also hunt bridges connecting 3+ of your seeds — they open new segments fast.
• Letting the map go stale. Fix: 10–15 minute weekly refresh with Last Verified dates.

1-week action plan

1. Day 1: Normalize names and run the scoring prompt. Save the top 10 per company.
2. Day 2: Verify the top 10; enforce the two-signal rule; update dates and confidence.
3. Day 3: Identify hubs and bridges; select 3 targets for immediate outreach.
4. Day 4: Draft and send 3 tailored outreach emails using signal-based angles.
5. Day 5: Build a one-page visual of High-confidence nodes; share with your team for alignment.
6. Day 6: Set a weekly 15-minute calendar block; note gaps to investigate next (e.g., missing suppliers or channels).
7. Day 7: Review metrics; adjust the scoring rubric if false positives are high.

Do this once and you’ll get clarity. Do it weekly and you’ll own the ecosystem narrative in your market. Your move.

Cut the busywork: use AI to decide what you keep, delegate, or refuse — and get tangible hours back this week.

Problem: Everything landing in your inbox looks important. You end up doing low-impact tasks that steal time from strategic work.

Why this matters: Every hour on low-value work is an hour not spent on revenue, relationships, or decisions that scale. Delegating right increases capacity, reduces mistakes, and lowers stress.

My short lesson: Use a fast, repeatable triage process (impact × effort) and let AI provide consistent recommendations: KEEP / DELEGATE / SAY NO. You stay the decider — AI gives unemotional, structured choices.

What you’ll need

• List of 20–40 tasks (email actions, meetings, admin, short projects)
• Estimated time per task (10–60 minute buckets)
• Conversational AI (chat tool) and a notes app or spreadsheet
• Simple SOP template: 2 lines (steps + acceptance criteria)

How to do it — step-by-step

1. Collect one week of tasks (30–60 minutes).
2. For each task add: time estimate, desired outcome, any constraints.
3. Paste tasks into the AI prompt below. Ask for: label (KEEP / DELEGATE / SAY NO), suggested role, 2-line delegation brief, acceptance criterion, tools/permissions.
4. Review AI output quickly. Accept or tweak top 10 delegatable items and assign owners with deadlines.
5. Create 2-line SOPs for repeated items and schedule a 15-minute weekly QA to review quality.

Copy-paste AI prompt (primary)

“You are my operations advisor. I will paste a numbered list of tasks with estimated time and desired outcome. For each task return: 1) one label: KEEP / DELEGATE / SAY NO, 2) suggested job title or role to assign to, 3) a 2-line delegation brief (steps + expected outcome), 4) one acceptance criterion, 5) any tools or permissions required. Use concise bullets. Tasks: [paste tasks here].”

Prompt variant (include risk)

“Same as above, but also include one risk/dependency per task and a 1-sentence mitigation.”

What to expect

• AI gives consistent triage in minutes. Use it as recommendation — not law.
• Expect 30–60% of routine tasks to be delegatable in month one.
• Initial time investment: 2–3 hours; weekly maintenance: 15–30 minutes.

Metrics to track

• % tasks delegated (target 30–60% first month)
• Hours freed per week
• Time-to-completion for delegated tasks
• Rate of rework / corrections
• Hours spent on strategic work (before vs after)

Mistakes & fixes

• Over-delegating without standards — Fix: attach a 2-line SOP and acceptance criterion.
• Vague task outcomes — Fix: add one clear expected outcome and deadline.
• Skipping review — Fix: 15-minute weekly QA; require assignee to log completion time and outcome.

7-day action plan

1. Day 1: Export tasks — 30–45 minutes.
2. Day 2: Run AI prompt and review labels — 45–60 minutes.
3. Day 3: Assign top 10 delegatable tasks with 2-line SOPs — 30 minutes.
4. Day 4: Communicate owners, deadlines, and acceptance criteria — 30 minutes.
5. Day 5–7: Monitor progress, collect hours saved, tweak prompts/SOPs.

Start small: one week of tasks, one review per week, measure hours freed, then scale.

Your move.

— Aaron

Hook: Turn messy feedback into a PMF Gap Report in 90 minutes. Not just sentiment — a ranked list of revenue-weighted problems with ready-to-run experiments.

The problem: Sentiment alone is blunt. Without weighting by customer value, recency and where feedback sits in the journey, you’ll chase noise and miss levers that move conversion and retention.

Why it matters: You need fast, defensible priorities that tie to revenue at risk and activation blockers. Do that and your roadmap shifts from opinion to impact.

What I’ve learned: Layer AI summaries with a simple scoring model and a two-step validation. The win is speed plus signal quality — fewer detours, faster movement on activation, retention and expansion.

What you’ll need

• A CSV of comments (200–1,000 rows) with columns: id, comment, channel, user_type (trial/paid), plan_tier, mrr (or proxy), tenure_days, date, product_area (optional).
• An AI tool that can follow structured prompts.
• 15 minutes to prep data; 45–60 minutes to run and review outputs; 30 minutes to decide experiments.

Insider upgrade: the PMF Gap Score

• PMF_Gap = Share_of_mentions × Negative_rate × Value_weight × Recency_weight × Journey_weight.
• Value_weight: trial=0.7, paid=1.0, enterprise/high MRR=1.5.
• Recency_weight: last 30 days=1.2, older=1.0.
• Journey_weight: onboarding/core path=1.3, peripheral=1.0.

Step-by-step (do this)

1. Prep your sheet (15 min). Dedupe comments, fill user_type, plan_tier, mrr (estimate if needed), tenure_days and product_area. Keep at least 200 rows. Expect a clean file AI can parse.
2. Two-pass tagging (20–30 min). First pass: AI proposes themes (6–12 max) and JTBD (“job-to-be-done”) labels. Second pass: you merge near-duplicates (e.g., “speed” and “performance”). Expect a stable theme set.
3. Score with weights (10–15 min). Ask AI to compute Negative_rate, Count, Share_of_mentions and PMF_Gap using the weights above. Expect a ranked list by PMF_Gap.
4. Extract customer language (5–10 min). For top 5 themes, pull 3 verbatim quotes each. Expect crisp, copy-ready phrasing for UX and messaging.
5. Convert to experiments (15–20 min). For each top theme, generate 3 small experiments: change, metric, target uplift, effort. Expect one low-effort winner to ship this sprint.
6. Validate quickly (same day). Micro-survey one question to affected users: “How much does [issue] block your goal?” Scale 1–5. Expect confirm/reject signals before dev time is committed.

Copy-paste AI prompt (robust, use as-is)

“You are a product analyst. You will receive CSV-like rows with columns: id, comment, channel, user_type (trial/paid), plan_tier, mrr, tenure_days, date, product_area. Tasks: 1) Propose a concise theme for each comment and a JTBD (job-to-be-done) phrase. 2) Aggregate by theme and return for each theme: a) 2–3 sentence summary in plain language, b) total count, c) avg sentiment from 0 (very negative) to 1 (very positive), d) negative_rate (share of comments with sentiment <0.4), e) share_of_mentions (count / total rows), f) value_weight (trial=0.7, paid=1.0; if plan_tier indicates enterprise or mrr>=$500 then 1.5), g) recency_weight (date within last 30 days=1.2 else 1.0), h) journey_weight (if product_area includes onboarding, signup, activation, core feature → 1.3 else 1.0), i) PMF_Gap = share_of_mentions × negative_rate × value_weight × recency_weight × journey_weight. 3) Return a ranked list of themes by PMF_Gap (highest first). 4) For the top 5 themes, provide: three representative quotes (verbatim), and three lean experiments with: hypothesis, change description (under 140 chars), primary metric (e.g., activation rate, time-to-first-value, conversion to paid), expected effect size (1–5%), effort (S/M/L), and success threshold. Output as plain text sections per theme with the fields clearly labeled. Also output a 5-line executive summary at the top with the top 3 themes and their PMF_Gap scores.”

What to expect from the prompt: a one-page executive summary, a ranked theme list, a quote bank you can paste into tickets, and 15 experiment ideas with success thresholds. If the AI returns more than 12 themes, re-run asking to cap themes at 10 and merge similar labels.

Metrics to track

• PMF_Gap top theme score (target: down 30% in 30 days).
• Activation: percent reaching first value within 24/72 hours (target: +3–5% absolute).
• Trial → paid conversion (target: +1–2% absolute over two sprints).
• Support tickets per 100 new users on top theme (target: −20%).
• MRR at risk: sum of MRR tied to users mentioning a top theme (target: −15%).

Common mistakes & fixes

• Too many themes — Fix: cap at 10 and force merges; otherwise prioritization dilutes.
• No denominator — Fix: always use share_of_mentions, not just counts.
• Chasing old issues — Fix: use recency_weight; archive anything older than 90 days unless high value.
• Ignoring value — Fix: include plan_tier/mrr; don’t let free-user noise steer roadmap.
• Vague experiments — Fix: require metric, effect size and threshold before work starts.

1-week action plan

1. Day 1: Clean the CSV; run the prompt; cap themes at 10; publish the ranked list and executive summary.
2. Day 2: Review top 5 themes; pick 3 experiments with S/M effort; define metrics and thresholds.
3. Day 3: Launch a 1-question micro-survey to users who mentioned the top theme; book 3–5 quick calls.
4. Day 4: Ship one low-effort experiment (e.g., onboarding copy, default setting, loading state).
5. Day 5: Monitor activation and tickets; re-run the prompt on new comments (incremental).
6. Day 6–7: Review early results vs thresholds; greenlight next experiment or roll back. Update PMF_Gap trend.

Pro tip: Run prompts in two passes — discovery (open themes) then convergence (cap themes, merge labels, compute PMF_Gap). This avoids the “loud outlier” trap and stabilizes priorities.

Your move.

Hook: Use AI to speed skin retouching, not to erase reality. The goal: cleaner skin that still reads human at 100% and on a phone.

The problem: Most people over-smooth, kill pores and lose midtone texture. That’s what makes portraits look “plastic” and wrecks client trust.

Why it matters: Natural retouching reduces revision requests, speeds delivery and keeps a consistent look across a shoot — all measurable in client satisfaction and turnaround time.

Core lesson from practice: Work in small, repeatable passes, non-destructively. Use masks for local control and measure impact at 100% and at final export sizes.

Quick checklist — Do / Don’t

• Do: keep originals and work on virtual copies or layers.
• Do: preserve texture (target 60–80% retention).
• Do: correct exposure and white balance before retouching.
• Don’t: apply a single heavy global smoothing slider.
• Don’t: remove all pores, fine lines or natural specular highlights.
• Don’t: skip 100% checks and device previews.

What you’ll need

• Raw or high-res JPG file.
• An AI retouch tool/plugin with masks and a detail/texture slider.
• A consistent screen and time to inspect at 100%.

Step-by-step routine (how to do it)

1. Global corrections: exposure, white balance, tint and gentle color grading to get skin tones right.
2. Create a duplicate layer/virtual copy; work there so you can revert quickly.
3. Run AI skin pass at low strength (20–35%). Set texture/detail retention to ~65–80%.
4. Paint masks for targeted work: under-eyes, redness, jawline. Use lower strength for under-eye vs cheeks.
5. Add +5 to +12 local clarity/micro-contrast to bring back midtones and pores where needed.
6. Sharpen for final output and export proofs at web and print sizes; check on phone and monitor.

Copy-paste AI prompt (use as-is)

“You are an expert portrait retoucher. Reduce visible blemishes and even skin tone while preserving natural skin texture (retain about 70% texture). Remove isolated spots and stray hairs. Soften under-eye shadows slightly without erasing fine lines. Maintain natural highlights and pores; avoid any plastic or overly smooth appearance. Deliver a subtle, natural portrait ready for web and print.”

Worked example

Shot: soft window light, RAW. Global: +0.2 EV, warm WB +200K. AI skin pass: 30% strength, texture 70%. Mask under-eyes: lower contrast by 8% only. Add +8 clarity to face layer. Inspect at 100% — if cheeks look glassy, drop AI strength on that mask by 10%.

Metrics to track (KPIs)

• Time per image (target: 5–15 minutes).
• Revision requests (%) from clients.
• Consistency score across set (visual checklist: tone match, texture parity).
• Acceptance rate of first proofs.

Mistakes & fixes

• Plastic look: lower smoothing, raise texture slider, add local clarity +5.
• Dull eyes: dodge iris highlights, add tiny sharpen to eyes layer.
• Inconsistent set: create a reference edit and batch-apply base adjustments, then refine locally.

1-week action plan (practical)

1. Day 1: Run the 20-minute test on one portrait using the routine above; export web+print proofs.
2. Day 2–3: Process 5 more images from the same shoot, apply the reference edit, note time and adjustments.
3. Day 4–5: Tweak your default AI strength/texture based on results; document settings.
4. Day 6: Create a one-page cheat sheet with your go-to values (strength, texture, clarity ranges).
5. Day 7: Deliver a short proof set to a client or peer, collect feedback and record revision rate.

Your move.

Hook: Lock the palette, then everything else. Color drift is wasted time and rework.

Problem: Midjourney loves to “improve” color with lighting and gradients. That’s artful, not brand-safe.

Why it matters: Consistent palette = faster approvals, fewer edits, repeatable assets across campaigns.

Lesson from the trenches: Show the palette visually first, then state it plainly. Weight the palette higher than the style words, and keep the model’s stylization low. Small constraints, big control.

• Do
  • Upload a clean 1:1 swatch image (3–6 solid color blocks, white background) and attach it first.
  • List exact hex codes and say “limited palette — use only these colors.”
  • Weight the swatch higher using multi-prompt: swatch ::3, text ::1.
  • Use –style raw and a low stylize value (e.g., –s 50–100) to reduce “art direction.”
  • Set a seed for repeatability (e.g., –seed 123), and keep –chaos modest (0–10).
  • Constrain look: “flat colors, no gradients, no textures, neutral lighting.”
  • Start with 3–5 core colors; add accents later if needed.
• Do not
  • Use vague color words (“warm,” “vibrant,” “moody”).
  • Overload adjectives (it dilutes the color constraint).
  • Run high stylize or high chaos if you need brand accuracy.
  • Expect photographic scenes to match on the first try—iterate and nudge.

1. What you’ll need
   • Midjourney on Discord.
   • 3–6 hex colors you must keep.
   • 1:1 palette swatch image (no text, just blocks).
2. How to do it (step-by-step)
   1. Build a swatch: 800×800px, 3–6 equal squares, white background. Save as a single image.
   2. In Discord, start your prompt and attach the swatch first. Then write short subject/style text.
   3. List hex codes and say “limited palette — use only these colors.” Add: “flat colors, no gradients, no textures, neutral lighting.”
   4. Control creativity: add –style raw –s 50–100 –chaos 0–10 –seed [number] –ar [ratio].
   5. Generate 4 variations, pick the closest, request variations on that one, then upscale. Use Vary (Region) to fix strays.
   6. If color drifts: reduce adjectives, drop to 3 core colors, increase swatch weight to ::3 or ::4, and rerun.
3. What to expect
   • First pass: ~50–75% palette adherence for simple/graphic styles.
   • After 1–3 iterations with swatch + hex: high adherence. Photo-like scenes may need light post-editing.

Worked example (copy-paste prompts)

Assume palette: #0B132B, #1C2541, #5BC0BE, #FDECEF

Poster (flat graphic)Attach your swatch image first, then paste:

[swatch attachment] ::3 minimalist travel poster, bold geometric skyline, high contrast ::1 limited palette — use only these colors: #0B132B, #1C2541, #5BC0BE, #FDECEF, flat colors, no gradients, no textures, centered composition, clean negative space —style raw —s 75 —chaos 6 —ar 3:4 —seed 123

Product (clean studio)Attach the same swatch, then paste:

[swatch attachment] ::2 sleek skincare bottle on seamless backdrop ::1 limited palette — use only these colors: #0B132B, #1C2541, #5BC0BE, #FDECEF, neutral white balance, soft shadow, no colored lights, no reflections, no gradients —style raw —s 60 —chaos 4 —ar 4:5 —seed 123

Social tile (brand quote card)Attach the swatch, then paste:

[swatch attachment] ::3 minimalist square quote card, bold typographic layout ::1 limited palette — use only these colors: #0B132B, #1C2541, #5BC0BE, #FDECEF, flat colors, no textures, strong hierarchy, generous margins —style raw —s 70 —chaos 5 —ar 1:1 —seed 123

Metrics to track

• Palette adherence rate: % of images using ≥80% listed hexes (target 80%+ by iteration 2).
• Iterations to approval: aim for ≤3 per asset.
• Time to final: start-to-export in minutes (target ≤12 for flat graphics, ≤20 for product).
• Post-edit minutes: keep color-correction under 3 minutes/image.

Common mistakes and fast fixes

• Too many descriptors causing drift → Strip to subject + palette + constraints.
• Six+ colors fighting for space → Start with 3–4; add accents in a second pass.
• High stylize or chaos → Drop to –s 50–100, –chaos ≤10, use –style raw.
• Palette image ignored → Weight it higher (::3/::4), attach first, and reduce text weight.
• Photo scenes look off-tone → Use “neutral lighting, accurate white balance, no colored gels,” and expect a minor post nudge.

1-week action plan

1. Day 1: Finalize 3–6 hex colors. Build a clean 1:1 swatch.
2. Day 2: Run the poster prompt (4 variations). Log adherence and time.
3. Day 3: Run the product prompt. Adjust –s and swatch weight based on results.
4. Day 4: Run the social tile. Capture best-performing settings.
5. Day 5: Iterate once on each asset, fix strays with Vary (Region), export.
6. Day 6: Light post color-correct where needed (≤3 minutes each).
7. Day 7: Review metrics, lock a “house” prompt with your winning settings.

Paste your 3–6 hex codes and the asset type you care about most (poster, product, or social). I’ll tailor three prompts you can run today and set targets for adherence and time.

Your move.

Good call on the 15–30 minute weekly review. That’s the safety rail that keeps automation tight. Let’s layer in two moves that materially cut admin time and tighten cash flow.

Quick win (under 5 minutes): turn on your accounting app’s payment link on the default invoice template and set terms to Net 7. Then add one extra reminder at 21 days overdue with a firm, friendly note. Expect faster pays and fewer chases.

The real bottleneck

You’re losing time in three places: 1) processor payouts (Stripe/PayPal) not matching cleanly, 2) receipts lingering in email, 3) inconsistent categories and rules. Fix these and you’ll automate 60–80% of the workload without losing control.

Insider lesson

Treat your payment processor like a bank account in your accounting app. Record sales gross, record fees separately, and reconcile the net payout as a transfer. That single model eliminates double-counting and makes month-end painless.

What you’ll need

• Accounting app: QuickBooks Online, Xero, FreshBooks, or Wave.
• Payment processor: Stripe or PayPal (connect the native feed if offered).
• Receipt OCR: built-in mobile app, Hubdoc, or Dext.
• Automation: Zapier or Make.
• One Google Sheet for a lightweight backup ledger.

Step-by-step: 60-minute foundation

1. Connect bank + processor feeds. In your accounting app, connect your business bank and card. Then connect Stripe/PayPal as its own “bank account.” Expect historic data to pull in within hours.
2. Create 7 categories (simple beats perfect): Income, Cost of Goods/Materials, Contractors, Software/Subscriptions, Meals/Travel, Bank/Processor Fees, Misc.
3. Set 5 bank rules (start narrow):
   1. Stripe payouts → Transfer to Checking (Description contains “Stripe Payout”).
   2. Stripe/PayPal fees → Bank/Processor Fees (Payee contains “Stripe” or “PayPal” and small negative amounts).
   3. Adobe/Canva → Software/Subscriptions.
   4. Uber/Taxi → Meals/Travel.
   5. Your domain host → Software/Subscriptions.
4. Invoice template: add payment link, Net 7 terms, and auto-reminders at 7 days before due, 7 days overdue, and 21 days overdue. Use a subject line with invoice number for easy search.
5. Receipt capture: enable OCR. Forward receipts to your OCR inbox or snap photos. Set it to create draft expenses only (no autopublish yet).
6. Two automations:
   1. Emailed receipt → OCR → Draft expense in accounting app.
   2. Invoice paid → Append a row to Google Sheet (date, client, invoice #, gross, fees, net, payment method).
7. Close-the-loop test (10 minutes): issue a $10 invoice to yourself, pay it via Stripe. You should see: gross sale in the Stripe feed, fee as an expense, net payout transferred to checking, invoice marked paid automatically, and a new row in your Google Sheet.

Metrics to track (weekly)

• DSO (Days Sales Outstanding): target under 15 days for a side hustle.
• % auto-categorized: aim for 70%+ by week 2, 85%+ by week 4.
• Payout match rate (processor → bank): target 100% matched with zero unexplained balances.
• Time-to-invoice: issue within 24 hours of delivering work.
• Receipt lag: days between purchase and captured receipt; keep under 2 days.

Common mistakes & fixes

• Double-counting sales (recording both bank deposit and invoice payment manually). Fix: connect the processor feed, treat payouts as transfers, and let the app auto-match.
• Rules too broad (“contains ‘Inc’”). Fix: tighten with exact vendor names or amounts ranges; review the first 20 auto-categorized items.
• OCR autopublish on day one. Fix: keep drafts on for two weeks; approve in your weekly review, then enable autopublish for only trusted vendors.
• Unmatched refunds/chargebacks. Fix: create a rule to map processor refunds to a “Refunds/Contra Income” category and match to the original invoice where possible.
• Too many categories. Fix: collapse to the 7 above; add detail later if truly needed.

High-value template: payment processor clearing

• Connect Stripe/PayPal as a separate bank account in your accounting app.
• Sales are recorded gross in the processor account; fees hit Bank/Processor Fees.
• Payouts are recorded as transfers from Processor → Checking.
• Weekly check: processor account should trend near zero after payouts; any residual means unmatched fees, refunds, or timing differences.

Copy-paste AI prompt

“Act as a pragmatic small-business bookkeeper. I run a [service/product] side hustle in [country/currency]. I use [AccountingApp] with [Processor: Stripe/PayPal] and [BankName]. Build a simple automation plan that includes: (1) 7 chart-of-accounts categories, (2) 7 precise bank rules (text matches and amount patterns) including processor payouts/fees, (3) an invoice template with Net 7 terms, payment link, and 3 reminder messages (before due, 7 days overdue, 21 days overdue), (4) a Stripe/PayPal clearing-account setup so sales are recorded gross, fees separate, and payouts transfer to checking, (5) a Zapier flow for emailed receipts → draft expenses and invoice paid → Google Sheet log (provide exact field mapping), (6) a 15-minute weekly review checklist, and (7) KPI targets for DSO, % auto-categorized, payout match rate, and time-to-invoice. Present step-by-step instructions I can follow without accounting jargon.”

1-week action plan

1. Day 1: Connect bank + processor feeds; set Net 7 + payment link on invoice template; enable 3 reminders.
2. Day 2: Create the 7 categories; add 5 bank rules (fees, subscriptions, travel, materials, payouts).
3. Day 3: Turn on OCR; forward 10 receipts; keep as drafts.
4. Day 4: Build 2 automations (receipts → expense draft; paid invoice → Google Sheet).
5. Day 5: Run the $10 end-to-end test; confirm processor account clears to near zero after payout.
6. Day 6: First 15-minute review; fix mis-categories; tighten one rule.
7. Day 7: Capture baseline KPIs (DSO, % auto-categorized, payout match rate, time-to-invoice) and set next week’s targets.

Your move.

Good call on focusing on results and KPIs — that’s exactly where this starts and ends. AI can absolutely create product photos and mockups for your store, but only when you set clear requirements and measure outcomes.

The problem: Many sellers use AI images that look good in isolation but don’t drive clicks, conversions, or trust on product pages.

Why this matters: Product imagery is the top driver of click-through and purchase decisions online. Poor images cost traffic and sales; great optimized images increase conversion and reduce returns.

Experience / lesson: I’ve seen stores replace basic stock images with AI-generated, on-brand mockups and lift conversions by 10–35% within weeks by focusing on consistency, context, and clarity.

Do / Do not checklist

• Do provide exact product dimensions, brand colors, and 2–3 style references.
• Do generate a consistent set: white-background hero, 3 lifestyle shots, 1 scale/measurement shot.
• Do test variations against existing photos — A/B test.
• Do not rely on a single AI pass; iterate and curate selections.
• Do not use images that misrepresent product features or scale.

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

1. Gather inputs: clear product photos (top, side, in-hand if available), exact dimensions, logo file, brand color hex codes, example images you like.
2. Decide image set: hero (white bg), three lifestyle (indoor, outdoor, in-use), and one scale shot (with a common object or measurement overlay).
3. Use an AI image tool or generator and run 3–5 prompt variations per image type; choose the best 2 per type; refine for realism and consistency.
4. Post-process: align color, add brand logo/subtle watermark, export at web-optimized sizes (e.g., 2000px on the long edge, WebP/JPEG at 70–80% quality).
5. Upload and A/B test hero image vs. control for at least 2 weeks or 1,000 impressions per variant.

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

Create a high-resolution product photograph of a ceramic coffee mug for an e-commerce product page. Provide a clean white-background hero shot with soft, natural studio lighting, 45-degree angle, visible handle, true-to-life colors, natural shadow. Also create three lifestyle images: (1) mug on a wooden kitchen counter with morning light and coffee steam, (2) mug held in hand near a laptop, casual home office look, (3) mug on a picnic blanket outdoors. Include one scale shot next to a smartphone. Output 2000px long edge, realistic texture, no watermarks. Match color tones to HEX #6B3F2F (brand brown) for subtle accent elements like a coaster.

Metrics to track

• Product page conversion rate (before vs after)
• Click-through rate from category pages
• Add-to-cart rate
• Return rate (misrepresentation)
• Time to create / cost per image

Mistakes & fixes

• Blurry or inconsistent lighting —> fix by constraining style and using reference images.
• Wrong scale —> include scale objects or measurement overlays in prompts.
• Brand mismatch —> lock colors and logos during post-production.

One-week action plan

1. Day 1: Collect inputs (product specs, reference images).
2. Day 2–3: Generate 3 prompt variants per image type; review and select.
3. Day 4: Post-process selected images and export web sizes.
4. Day 5: Implement images on product page; set up A/B test.
5. Day 6–7: Monitor performance and note early signals (CTR, add-to-cart).

Worked example: For a ceramic mug, I’d create a white-background hero, two lifestyle shots (kitchen and desk), and a phone-scale shot. After A/B testing, prioritize the lifestyle image that yields higher CTR; swap hero if conversion improves.

Your move.

Good point — the 3-bullet follow-up + a bite-sized 30-day pilot is the fastest path from a one-off call to a recurring retainer. That single move reduces friction and gives clients an easy yes.

The problem: too many consultants treat calls as isolated events. Without a fast, measurable next step, prospects drift and you lose predictable revenue.

Why it matters: turning advisory time into recurring revenue multiplies your value-per-hour, stabilizes cashflow and makes growth scalable. You don’t need bigger clients — you need repeatable, predictable conversions.

Quick lesson: clients buy demonstrated outcomes, not promises. Deliver one measurable win in 30 days and retainers follow.

What you’ll need

• Call notes (3 bullets: main problem, immediate win, recommended next step).
• A simple 30-day pilot offer (weekly deliverables, 30-min check-ins, price band).
• Calendar link and two follow-up email templates.
• An AI writing assistant to draft and personalize outreach fast.

Step-by-step (do this now)

1. Within 24 hours: send a 3-bullet follow-up (value, quick win, clear next step).
2. Same day: use the AI prompt below to draft a 30-day pilot + onboarding checklist.
3. Day 2: send the short pilot proposal (one page): weekly deliverable, success metric, cadence, price band, start date, cancel-after-30-days option.
4. Automate reminders at day 3 and day 7; call if no response after 7 days.
5. If pilot accepted: schedule kickoff, deliver week 1 win within 7 days, measure and report progress weekly.

Metrics to track

• Follow-up sent within 24h rate (target: 95%).
• Pilot acceptance rate from follow-ups (target: 25–40%).
• Pilot-to-retainer conversion (target: 50%+).
• Time-to-first-value (target: 7 days).
• MRR per converted client and churn at month 2–3.

Common mistakes & fixes

• Waiting more than 24 hours — Fix: send follow-up before you close your laptop.
• Overloaded proposals — Fix: offer a single measurable success metric for 30 days.
• No calendar link — Fix: include one-click scheduling in every proposal.

7-day action plan

1. Day 1: Send the 3-bullet follow-up and AI-draft the pilot.
2. Day 2: Send the 30-day pilot proposal with calendar link.
3. Day 3: Reminder email if no reply.
4. Day 4–6: Prepare onboarding checklist and week-1 deliverable.
5. Day 7: Call if still no response or confirm kickoff if accepted.

Copy-paste AI prompt (use as-is)

“You are an expert business consultant. I just had a [length]-minute call about [topic]. Notes: [paste 3 bullets]. Create: 1) a 3-bullet follow-up email that highlights immediate value and asks for a 30-minute next meeting; 2) a short 30-day pilot proposal (weekly deliverables, meeting cadence, one clear success metric, price band, and a simple 30-day cancellation option); 3) a one-page onboarding checklist for week 1. Keep language simple and client-focused.”

Your move.

Aaron Agius