AI Business Transformation Strategy for Russian Hill Handyman (Q3–Q4 2026) Business context used: Rick Haelig is a one-person handyman in San Francisco, charging $150/hr for in-person visits. Leads come via phone, text, and WordPress form. Implementation capacity is 5–10 hours/week. Rollout target is Q3–Q4 2026. Deliverable 1 — AI Transformation Roadmap (Markdown) Executive Summary Rick’s core model is commercially sound: convert low-value, repetitive calls into a $75 DIY Video Guide product, then apply that $75 as a credit if the customer later books an in-person visit at $150/hr. This creates three strong outcomes: (1) paid value capture on small requests that usually consume unpaid triage time, (2) better client experience through immediate self-service, and (3) cleaner upsell path to paid field work without perceived double-charging. For a solo operator, this is especially attractive because it monetizes demand that would otherwise clog the calendar. The model’s structural weaknesses are manageable with guardrails. Biggest risks are mis-triage (unsafe DIY advice), liability from imperfect AI visuals, credit/refund disputes, and operational overload from content production. The strategy below keeps Rick protected by: strict exclusion of permit-sensitive/high-risk topics, mandatory human review of safety-critical video steps, a clear legal disclaimer embedded in every video, and a simple credit ledger tied to Jobber records. With disciplined scope, this can be launched within Rick’s 5–10 hours/week bandwidth. Model Guardrails (Non-Negotiable) DIY-eligible only for low-risk tasks (e.g., faucet aerator/cartridge swaps, toilet flapper replacement, cabinet hinge adjustment, weatherstripping, basic caulking). Hard exclusions: electrical panel work, gas line work, structural changes, permit-required work, and any task where shutoff/isolation is unclear. Compliance boundary (California): Unlicensed handyman cap: $1,000/project (labor + materials). SB 779 minimum penalties: $1,500 for unlicensed violations. DIY video is sold as a digital educational product, not a construction contract. Safety layer: every AI-generated instructional video gets 5–10 minute human safety review before publishing. Credit policy clarity: one credit per qualifying video purchase, applied to same-job first invoice line item; timestamped in ledger to prevent disputes. Phase 1 — Intake Triage Automation (Month 1–2: Jul–Aug 2026) Goal: Intercept inbound requests and route each lead to either DIY checkout or direct booking. Workflow (phone/text-first): Client contacts Rick by phone or SMS. If call is missed, auto-SMS reply from Twilio: “Thanks for contacting Russian Hill Handyman. I can help triage your request and send a DIY option or book a visit.” Twilio forwards message into GPT-4o (or Gemini) triage prompt. AI classifies request: DIY-eligible Requires Rick visit Excluded/high-risk (forced direct booking or decline with safety note) Concrete faucet example: Client text: “My kitchen faucet is dripping.” AI asks 2 qualifying questions: “Is your faucet single-handle or double-handle?” “Can you locate and turn off the under-sink shutoff valves?” If answers indicate low-risk DIY candidate → send $75 video checkout link. If uncertain/safety concern → offer direct booking at $150/hr. WordPress form mirroring: WordPress form submission → Zapier/Make trigger → same triage logic → auto email/SMS follow-up with either: DIY checkout link, or booking link / callback scheduling. Rick time commitment: ~4–6 hours/week for first 2 weeks, then 2–3 hours/week tuning prompts and fallback responses. Phase 1 outputs: Triage scripts and routing rules FAQ response templates Basic audit log (message, classification, outcome) Compliance integration: triage prompt must detect permit-sensitive keywords and force non-DIY handling. Phase 2 — DIY Video Content Engine (Month 2–4: Aug–Oct 2026) Goal: Build a 15–20 video library for highest-frequency, lowest-risk requests. Recommended production stack: Script and shot outline (Rick-authored from real jobs) AI draft visuals via Runway Gen-3 Alpha, Pika 2.0, or OpenAI Sora API Rick voice-over (or text narration) Final human safety review on critical steps (5–10 minutes/video) Why hybrid (AI + human): AI video still has hallucination risk (wrong part/tool/procedure). Early-2026 models still show temporal drift/mechanical inconsistency across frames. For instructional repairs, Rick’s short review is mandatory for brand and safety protection. Mandatory disclaimer (embed in every video and checkout page): "This AI-assisted DIY video is for informational purposes only. Russian Hill Handyman is not liable for user errors, property damage, or injuries. For complex repairs, professional service is strongly recommended." Content priority order (first 6–8 videos): Dripping faucet basics Running toilet/flapper replacement Clogged P-trap cleaning Loose cabinet hinge realignment Door strike plate adjustment Re-caulking tub/sink seams Rick time commitment: ~3–5 hours/week (batch scripting + weekly review session). Compliance integration: no videos on excluded permit/high-risk categories. Phase 3 — Monetization & Credit-Back Infrastructure (Month 3–5: Sep–Nov 2026) Goal: Turn DIY guidance into a clean, trackable revenue stream with seamless escalation. Checkout options: Option A: WooCommerce (WordPress-native) Best if Rick wants product catalog, coupon logic, and native site control. Option B: Stripe Payment Links (lighter lift) Best if Rick wants fastest launch with minimal site complexity. Credit-back implementation: Customer buys $75 DIY guide (auto receipt + job reference ID). If customer later books Rick for same issue: apply $75 credit against first invoice hour. at $150 first hour, net additional billed = $75 (no double-billing). Track credits in Notion or Airtable ledger: Customer Date/video purchased Job category Credit status (unused/applied/expired if policy used) System of record: Use Jobber for scheduling, quotes, invoicing, and client history. Add a custom field/tag: “DIY Credit Eligible / Applied”. Rick time commitment: ~2–4 hours/week during setup; ~30–60 min/week ongoing admin. Compliance integration: keep invoices/job scopes explicit; maintain clear line between digital education sale and in-person labor contract. Key Risks and Mitigations (Operational) Risk Impact Mitigation Mis-triage of unsafe jobs Safety/liability exposure Hard exclusion keywords + forced manual review branch AI video inaccuracies Bad instructions, reputation loss 5–10 min human review on safety-critical steps Credit disputes Revenue leakage, client friction Written credit policy + ledger + invoice line item transparency Scope creep for solo operator Burnout, inconsistent delivery Limit to top 15–20 videos, batch work weekly Regulatory missteps Penalties/compliance risk Enforce $1,000 project cap awareness; no permit-sensitive tutorials Deliverable 2 — Client Journey Flowchart (ASCII/Text Diagram) [Client contacts Russian Hill Handyman] -- via Phone -- via Text -- via WordPress Form v [AI Triage Agent: Twilio + GPT-4o/Gemini (+ Zapier/Make for web form)] Note: asks 2–3 qualifying questions What exact fixture/type is involved? Can shutoff/isolation be done safely? Any signs this is permit-sensitive/high-risk? v +-------------------- Decision --------------------+ DIY-eligible Complex / High-risk / Permit-sensitive v v [Send $75 DIY Video Checkout Link] [Book Rick directly at $150/hr] v v [Client pays $75] [Standard service workflow] v +------------------ Outcome ------------------+ A1: Client succeeds with video A2: Client needs help anyway v v [Job done remotely] [Book in-person visit] [Rick earns passive $75] [Apply $75 credit to first hour] [First hour $150 - $75 credit = $75 net additional] Deliverable 3 — AI Tool Matrix (Markdown Table) Ease of Setup scale: 1 = hardest, 5 = easiest. Tool Name Category Best Use Case for Rick Price Tier (2026) Ease of Setup (1–5) Key Risk or Limitation Runway Gen-3 Alpha AI Video Generation Draft visuals for common DIY repair scenarios from scripts Medium 3 Temporal drift and tool-motion inconsistency; must human-review Pika 2.0 AI Video Generation Fast short-form explainer clips and variations Low–Medium 4 Visual realism may vary; can imply incorrect technique OpenAI Sora API AI Video Generation (API) Programmatic generation pipeline for repeatable templates Medium–High (usage-based) 2 Integration complexity + still imperfect mechanical continuity Twilio + GPT-4o SMS/Phone AI Triage Auto-response, intake questions, DIY vs booking routing Medium (usage-based) 3 Misclassification risk without strict prompts/guardrails Twilio + Gemini (alternative backend) SMS/Phone AI Triage Alternate model stack for routing and response cost/performance tuning Medium (usage-based) 3 Output variability across models; requires prompt QA Make.com Workflow Automation Connect WordPress forms to triage, CRM, and notifications Low–Medium 4 Scenario complexity can become brittle over time Zapier Workflow Automation Fast no-code form-to-SMS/email routing Medium 5 Higher cost at scale; multi-step logic can get expensive WooCommerce WordPress Checkout Full DIY product storefront on Rick’s site Low–Medium 3 Plugin overhead/maintenance; checkout tuning needed Stripe Payment Links Lightweight Checkout Fastest path to selling $75 video guides Low 5 Less native catalog/upsell control than full WooCommerce Jobber CRM + Scheduling + Invoicing System of record for jobs, invoices, and credit application notes Medium 4 Requires process discipline for consistent tagging Notion Credit Ledger Tracking Simple manual credit tracking with lightweight database Low 5 Manual entry errors if not tied to routine Airtable Credit Ledger Tracking Structured credit ledger with filters/views/automation Low–Medium 4 Slightly steeper setup than Notion Critical note: For AI video tools, temporal drift and mechanical inconsistency remain known limitations in early 2026. Rick’s 5–10 minute human safety review is non-negotiable before publishing any instructional content. Top 3 Quick Wins (This Week, Each Under 2 Hours) Publish the credit-back policy on the website (60–90 min). Add a clear FAQ block: “$75 DIY guide is credited to your first in-person hour for the same job.” Set up one Stripe Payment Link for a pilot DIY guide (30–45 min). Create product: “Fix a Dripping Faucet — DIY Video Guide ($75)” and place link/button on WordPress. Draft triage script + 3 qualifying question templates (60–90 min). Prepare copy for SMS replies and WordPress follow-up automation so implementation can start immediately in Twilio/Zapier/Make.