Ops Runbook

Ops Runbook — AI Opportunity Audit Delivery

The engagement timeline, QC checklist, welcome packet, GHL pipeline, and scope-creep defense.

Purpose: So nothing slips between “client paid” and “client raves.” This is the operational layer that makes you look like a pro, not a person with a landing page.

---

1. ENGAGEMENT TIMELINE (day-by-day)

Three timelines by tier. Day 0 = the day intake is complete AND all interviews are done (the clock starts here, per the SOW Section 1.3).

Snapshot ($1,500) — 7 days

Day	Action	Owner	Mode
−3 to 0	Kickoff: send welcome packet + MNDA + intake form. Book 1 interview.	You	Semi
0	Intake complete + interview done → trigger n8n “Run Analysis”	n8n	Auto
1	Claude drafts scores, top-5 quick wins, 1-pg roadmap, ROI	n8n	Auto
1–2	You review drafts, edit, approve in Supabase	You	Manual
2	On approval → n8n generates PDF + Placid graphic + portal + Loom script draft	n8n	Auto
2–3	You record Loom (5 min, read/refine the draft script)	You	Manual
3	n8n sends delivery email (PDF + portal + Loom + debrief booking link)	n8n	Auto
4–7	Debrief call (45 min). Send recap + referral prompt.	You	Manual

Opportunity Audit ($3,000) — 14 days

Day	Action	Owner
−5 to 0	Welcome packet + MNDA + intake + book 2–3 interviews	You
0	All interviews done → trigger analysis	n8n
1–2	Claude drafts all 5 sections (scores, 15–20 opportunities, roadmap, ROI, Loom script)	n8n
2–4	You review + edit + approve	You
4–5	n8n renders deliverables	n8n
5–6	You record Loom (8 min)	You
6	Delivery email sent	n8n
7–14	Debrief (60 min). Recap + referral prompt + retainer pitch.	You

Deep Dive ($5,000) — 21 days

Day	Action	Owner
−7 to 0	Welcome + MNDA + DPA + intake + stakeholder roster + book 5–6 interviews	You
0	All interviews done → trigger analysis	n8n
1–3	Claude drafts all sections + governance/risk review	n8n
3–6	You review + edit + approve	You
6–7	Render + Loom	n8n + You
7	Delivery email	n8n
8–21	Executive debrief + board-ready 1-pager handoff. Retainer pitch.	You

Your time per engagement: ~3–5 hrs (Snapshot), ~5–7 hrs (Opportunity), ~8–10 hrs (Deep Dive). The rest is automated.

---

2. CLIENT WELCOME PACKET

Sent automatically the moment payment clears (n8n Step 0 → GHL email). The email links to a simple Framer or Notion “welcome page” containing:

2.1 Welcome email (send this)

Subject: You’re in — your AI Opportunity Audit starts now

Hi [FIRST],

Welcome to the AI Opportunity Audit. Here’s exactly what happens next, so you always know where we are.

3 things to do now (10 minutes):

1. Sign the engagement letter & MNDA → [link]
2. Complete the intake questionnaire → [Fillout link]
3. Book your interview(s) → [GHL calendar link]

The clock on your [7/14/21]-day delivery starts the day your intake is done and interviews are complete — so the faster these three, the faster your roadmap.

What you’ll get: your AI Readiness Score (0–100), a ranked opportunity matrix, a 30/60/90 roadmap, an ROI projection, a recorded walkthrough, and a live debrief call.

Your single point of contact: me. Reply to this email anytime.

— Dwad

2.2 Welcome page contents (Framer/Notion)

• What to expect — timeline, deliverables list, a sample (anonymized) report PDF.
• How to prepare for your interview — 1-page prep guide: come ready to talk about your top 5 most time-consuming tasks, your data systems, and where you feel the friction. No prep deck needed.
• Who to invite to the interview — role guidance (for Deep Dive: ops, finance, tech, frontline).
• Security & confidentiality — how your data is handled, where it’s stored, when it’s deleted. Link to the DPA.
• FAQ — the same FAQ as the sales page, plus delivery-specific questions.
• Single point of contact — your email + a backup channel (WhatsApp for urgent).

---

3. QC CHECKLIST (before anything goes to a client)

This is the most important operational artifact. Run it for every engagement. No exceptions.

3.1 Data & inputs

• Intake form 100% complete (no blank required fields)
• All scheduled interviews completed and transcribed
• Interview notes extracted and stored in Supabase
• Stakeholder names/roles spelled correctly in the data

3.2 Analysis output (Claude drafts — before you approve)

• All 7 dimensions have a score (1–5) and an evidence-based justification
• Binding constraints (2 lowest) correctly identified and flagged
• Overall score math is correct (weighted, binding-constraint method)
• 15–20 opportunities present (or top 5 for Snapshot), each with all fields populated
• Every opportunity is grounded in the client’s actual intake/interview data — no generic/hallucinated use cases
• Opportunities ranked by impact ÷ effort (descending)
• ROI math foots: hours × loaded cost × weeks × adoption % = stated savings
• Roadmap Phase 1 addresses the binding constraints
• No internal inconsistencies (e.g., a dimension scored 4 but justification says “nothing in place”)
• No client name / company errors

3.3 Deliverable rendering

• PDF renders without errors; all 10 pages present
• Every merge tag populated (no {client_name} left visible)
• Charts render (radar, scatter, bar, timeline) — no broken images
• Numbers in the exec summary match the detail pages
• Money slide ROI % calculated against the correct audit fee
• No text overflow / truncation / awkward page breaks
• Footer + confidentiality marking on every page
• Client logo + your logo present and correct
• Engagement ID + date correct

3.4 Shareable graphic + portal

• Placid graphic generated (both sizes); score, band, client logo correct
• Portal URL works; requires login; shows correct engagement data
• Loom recorded, links work, audio clear, no errors mentioned

3.5 Final human review (you, reading the PDF end to end)

• Read it as if you’re the client’s CFO — does the money slide hold up?
• Read it as if you’re the client’s CTO — are the technical claims defensible?
• Any claim you can’t personally stand behind → edit or remove
• Spelling, grammar, formatting final pass

Rule: If any item fails, fix before delivery. Never ship a deliverable with a known defect. The PDF is your reputation.

---

4. GHL PIPELINE STAGES

Set up this pipeline in GHL. Each stage triggers automations and gives you a live view of every engagement.

Stage	Trigger	Automation
1. Lead (scorecard complete)	ScoreApp webhook	Add tag scorecard_complete; send Email 1 (score + audit offer); tag hot leads 40–79
2. Qualified (scoping call booked)	Calendly/GHL booking	Send pre-call prep email; add to scoping call workflow
3. Proposal sent	You move manually	Auto-send proposal + SOW via GHL; set 3-day follow-up
4. Won (paid)	Stripe/GHL payment	Send welcome packet; create engagement in Supabase; move to In Progress
5. In Progress — Intake	Welcome sent	3-day reminder if intake incomplete; 7-day reminder; then flag
6. In Progress — Interviews	Intake complete	Reminders for each interviewee; reschedule flow
7. In Progress — Analysis	All interviews done	n8n trigger; you get “analysis ready” alert
8. In Progress — Review	Analysis drafted	You review + approve in Supabase
9. In Progress — Delivery prep	Approved	n8n renders; you record Loom
10. Delivered	Delivery email sent	Send debrief booking; +24h share-score email; +3d referral email
11. Debrief complete	You mark	Send recap + retainer pitch; move to Nurture
12. Retainer — pitched	You mark	Proposal auto-send; 7-day follow-up
13. Retainer — won	Payment	Create new engagement; credit audit fee
14. Nurture / referral	Auto	Quarterly check-ins; affiliate prompts

Custom fields to create in GHL: ai_readiness_score, ai_readiness_band, binding_constraint_1, binding_constraint_2, audit_tier, engagement_id, audit_status, referral_source, referral_link.

---

5. SCOPE-CREEP & DELIVERY-SLIP DEFENSE

5.1 The clock starts on intake, not on payment

SOW Section 1.3: the delivery timer starts when intake is complete AND interviews are done. This protects you from clients who pay, then take 3 weeks to fill the form and still expect 14-day delivery.

5.2 Intake deadlines

• Send the intake form with the welcome packet.
• 3 days with no response → auto-reminder email (GHL).
• 7 days → personal email: “Checking in — your delivery clock starts when this is done. Want help?”
• 14 days → call. If still stuck, offer a guided intake call (bill if it becomes consulting).

5.3 Interview no-shows

• Reminders 24h and 1h before (GHL).
• No-show → auto-offer 2 reschedule options within 5 business days.
• 2 no-shows → pause the engagement, personal note. You’re not a free calendar.

5.4 “Can you just also look at…?”

Every out-of-scope request routes through Change Control (SOW 1.5):

• “Happy to. That’s outside the current scope — I’ll send a Change Order with the added fee/timeline. Reply to approve and I’ll start.”
• Never start out-of-scope work without a written Change Order. This is the #1 margin killer for solo consultants.

5.5 “We want to add another department”

Same as above — Change Order. If it materially changes the engagement, re-quote or recommend a second engagement.

5.6 Data arrives late or incomplete

• If the client’s data is too incomplete to score defensibly, flag it: “Based on the information provided, here’s what we can conclude. To refine X and Y, we’d need [data]. We can add that as a Change Order, or proceed with stated assumptions.”
• Document the assumption in the report’s Appendix C (Limitations). This protects you and sets expectations.

5.7 The “can you make it faster” ask

• Snapshot is the fast tier. If they want Opportunity-level depth in Snapshot time → price is Opportunity, not Snapshot. State it plainly.

5.8 Refund & dispute handling

• All disputes go through SOW terms first, never straight to chargeback.
• If a client is unhappy, the debrief call is the resolution forum — fix it there before it escalates.
• “Found nothing” refund = full, no friction (it’s never happened, and a clean refund is better than a bad review).

---

6. INTERNAL DAILY/Weekly ROUTINE

Daily (10 min)

• Check GHL pipeline: any engagement stuck in a stage > expected time?
• Check n8n error logs: any failed runs in the last 24h?
• Respond to client emails (single batch, not all day).

Weekly (1 hr)

• Review all In Progress engagements; confirm next steps for each.
• Review ScoreApp responses; any hot leads (40–79) need personal outreach?
• Check referral activity; pay out any earned commissions via Wise.
• Update your benchmark data (median score, distribution) as responses accumulate.

Per engagement (during review)

• Run the QC checklist (Section 3).
• Update the Supabase status field as you progress.

---

7. CONTINUOUS IMPROVEMENT

• Log every defect — anything you caught in QC that Claude got wrong. Feed these back as prompt refinements (Asset 3 prompt library).
• Track time per engagement — actual vs. estimated. Adjust pricing when your time creeps up.
• Track conversion — scorecard → audit → retainer. Where does the funnel leak? Fix that stage.
• Collect a testimonial after every debrief — while they’re happy. Template: “In one sentence, what did this audit do for you?”
• Quarterly prompt review — re-read all 5 Claude prompts; update for new tools, new frameworks, lessons learned.

---

8. WHAT “DONE” LOOKS LIKE (definition of done per engagement)

An engagement is “done” when:

• All deliverables rendered and QC-passed
• Delivery email sent (PDF + portal + Loom + debrief link)
• Debrief call completed
• Recap email sent
• Referral link generated and prompted
• Testimonial requested
• Retainer pitch made (or explicitly declined)
• GHL contact updated to final stage
• Supabase engagement marked delivered
• Engagement folder archived (signed SOW, invoices, deliverables)

This is the bar. Hit it every time and you operate like a firm, not a freelancer.