The readiness audit produced the composite. The composite produced the band. The band sets the pacing. Now the owner needs the roadmap — a dated, named, metric-tied 12-month plan that sequences the AI bets, names the gating dependencies, and produces defensible numbers at each phase boundary. Twelve months. Four phases. Three months per phase. Each phase has a primary outcome metric, a vendor or tool stack, a named owner, and a hand-off to the next phase. Phase 1 (months 1-3): missed-call recovery plus AI call summaries via Avoca / Jobber AI Receptionist / Housecall Pro AI Agents plus ServiceTitan call-summary AI — lifts booking percent from 65% to 80%. Phase 2 (months 4-6): dispatch optimization plus tech scorecard cadence via Dispatch Pro / Sera / FieldEdge — lifts Revenue Per Truck 12-18%. Phase 3 (months 7-9): ride-along coaching plus AI-powered RC&D triage via Rilla, ResponsiBid, and ServiceTitan / Sera ticket classifiers — lifts close rate 8-14 points and cuts recall percent below 3%. Phase 4 (months 10-12): marketing attribution plus the owner's 12-metric dashboard via CallRail Conversation Intelligence, Ryze AI bid management, NiceJob, Podium AI Employee, Birdeye AI Employee, and Hatch — lifts GLSA ROAS 30-50% on top of a 3-4x baseline. This lesson is the roadmap — every phase, every metric, every dependency, every hand-off. The artifact every owner walks into the L4 capstone with and every PE partner expects on slide 4.

Why Four Phases of Three Months

The 12-month roadmap divides into four 90-day phases for three reasons. First, 90 days is the unit at which AI tool ROI becomes measurable — Avoca's case studies hit visible booking-percent lift in 60-90 days; Dispatch Pro's yield lift documents at 60 days; Rilla's close-rate lift consolidates at 60-90 days. Phase boundaries align to ROI visibility. Second, 90 days is the change-management unit the shop can absorb without breaking the team. Adding one major workflow tool per quarter is sustainable; adding three tools in one month produces tech pushback, CSR confusion, and owner overwhelm. Third, 90 days aligns to Nexstar, BDR, franchise-HQ, and PE board quarterly reviews — the roadmap fits naturally into existing reporting cadence rather than imposing new meetings.

Each phase produces one headline metric movement, two supporting metric lifts, and one operational artifact the team uses daily. Phase 1: booking percent (headline), missed-call percent and after-hours capture (supporting), the AI-summarized booking log the CSR floor reviews daily (artifact). Phase 2: RPT — Revenue Per Truck per day (headline), dispatch yield and show rate (supporting), the daily dispatch-board override log (artifact). Phase 3: close rate (headline), MPR — Membership Penetration Rate and recall percent (supporting), the weekly ride-along scorecard with AI-clustered coaching topics (artifact). Phase 4: GLSA ROAS (headline), cost-per-booked-call by source and RPL — Revenue Per Lead (supporting), the owner's 12-metric daily dashboard (artifact). By month 12, all four headline metrics have moved and all four artifacts are in daily rotation.

The phases also build on each other. Phase 1's clean booking data feeds Phase 2's dispatch optimization (Dispatch Pro can only route what the CSR floor booked clean). Phase 2's clean RPT data feeds Phase 3's coaching (Rilla's transcripts need accurate per-truck context). Phase 3's clean close and recall data feeds Phase 4's marketing attribution (Ryze AI's bidding model needs accurate closed-revenue signal). Skipping or shuffling phases breaks the substrate chain — buying Ryze AI before fixing CSR booking corrupts the bid optimization on bad closed-revenue signal. The sequence is not arbitrary.

Phase 1 (Months 1-3): The Missed-Call and Call-Summary Bet

The Phase 1 outcome target: booking percent lifts from the 65% industry baseline to 80% by end of month 3, missed-call percent drops from 22% baseline to under 8%, after-hours capture rate moves from 0-15% to above 60%. The lift drops $200K-$400K of recovered revenue at typical $5M shop economics over the 90-day window.

The tool stack: Avoca for in-platform-agnostic voice AI (the $1B-valued benchmark — Series B led by Meritech and General Catalyst in April 2026), or Jobber AI Receptionist for Jobber-stack shops at $99/mo or Plus plan, or Housecall Pro AI Agents for HCP shops, or ServiceTitan Voice (Titan Intelligence) for ServiceTitan-native shops. Plus ServiceTitan / Sera / HCP native call-summary AI that collapses CSR after-call work from 4 minutes to 15 seconds per call. The choice between bolt-on (Avoca) and in-software (ServiceTitan Voice, Jobber AI Receptionist, HCP AI Agents) is covered in L4 Chapter 2; for the readiness-audit-driven roadmap, the audit's CSR Maturity dimension drives the choice.

Phase 1 dependencies that gate launch: FSM Data Quality (Dimension 1 of the audit) at 12+ of 20 — Avoca's slot offer logic needs dispatch capacity data ServiceTitan publishes accurately; CSR Maturity (Dimension 2) at 9+ of 15 — the CSR floor must absorb transferred complex calls without dropping booking. Below either gate, Phase 1 launches month 2 or month 3 after a 30-60 day remediation. Above both gates, Phase 1 launches week one with a 90-day payback realistic on three of four bets.

Phase 1 named owners: the office manager or operations manager owns the daily Avoca review (4 p.m. CSR floor pull of AI-booked calls — verify slot, address, name, no fabricated promises); the marketing manager owns the missed-call recovery escalation list; the owner owns the weekly Phase 1 review on Friday at 4:30 p.m. with the CSR scorecard pulled. Month 1 milestone: Avoca live, 95% answer rate, daily review routine in place. Month 2 milestone: booking percent at 72%+, after-hours capture above 40%. Month 3 milestone: booking percent at 78-80%, missed-call below 10%, CSR ACW (after-call work) at 30 seconds average via ServiceTitan summary AI, 90-day Avoca ROI report ready for the Phase 2 launch meeting.

Phase 1 risk register: TCPA exposure on AI-voice consent (covered in L4 Chapter 5 — two-party-consent states require disclosure language in the AI's greeting; Avoca's 2026 build supports the disclosure); FTC review-bait exposure on AI booking confirmations (the AI cannot promise outcomes the shop has not delivered); CSR floor pushback at week 2 (covered in L4 Chapter 3 — Tech Pushback Playbook with comp plan and pilot script). The risk register travels with the roadmap into the L4 capstone defense.

Phase 2 (Months 4-6): The Dispatch and Scorecard Bet

The Phase 2 outcome target: Revenue Per Truck per day lifts 12-18% across the service truck fleet, dispatch yield improves on a per-call basis, show rate moves from 84% baseline to 92%+. At a 6-truck shop running $1,900/day average RPT pre-Phase-2, a 15% lift produces $285 incremental daily RPT per truck × 6 trucks × 250 working days = roughly $430K of incremental annual revenue, of which 60-70% drops to gross margin given marginal-truck-day economics.

The tool stack: ServiceTitan Dispatch Pro for ServiceTitan shops (the in-platform dispatch AI that re-evaluates the board every 10 minutes against revenue-per-truck targets, skill matching, geography, and capacity); Sera Systems' profit-aware AI scheduling for Sera shops (the 2026 dispatch AI that prioritizes margin per call); FieldEdge AI scheduling for FieldEdge shops; or the in-platform dispatch AI for Jobber and Housecall Pro shops. Add to this: the AI-summarized daily dispatch override report that surfaces dispatcher override patterns weekly for service manager review.

Phase 2 dependencies that gate launch: Tech Scorecard and Ride-Along Maturity (Dimension 3 of the audit) at 12+ of 20 — the service manager must have a working scorecard cadence to absorb the dispatch-yield-by-tech data; FSM Data Quality (Dimension 1) at 14+ of 20 — Dispatch Pro reads equipment, skill, geography, and ticket data that must be clean. Below the Dimension 3 gate, Phase 2 requires 30 days of service-manager training on the scorecard routine before the dispatch AI launches (typical 2-hour weekly training plus daily 15-minute scorecard review habit).

Phase 2 named owners: the dispatcher owns the daily override log with documented reason per override (target: dispatch AI suggestion accepted 85-90% of the time; overrides at 5-15% with logged reason); the service manager owns the weekly RPT-by-tech review and the dispatch-yield trend; the owner reads the dispatch-yield trend on Friday. Month 4 milestone: Dispatch Pro live, override discipline at 95%+ logged, baseline RPT measured per truck. Month 5 milestone: RPT trend up 6-10%, show rate at 89%+. Month 6 milestone: RPT lift at 12-18%, show rate at 92%+, dispatcher comfortable with the AI's daily board, service manager running the weekly scorecard rhythm.

Phase 2 risk register: dispatcher pushback in week 2-4 (the dispatcher feels the AI is overriding their judgment; comp plan and dispatch-override-discipline routine resolve it); tech complaints about routing changes (the AI's profit-aware routing sometimes moves a tech off a familiar zone; tech-floor communication and ride-along data resolve it); the Dimension 3 score must remain stable — service manager turnover during Phase 2 typically delays the RPT lift by 60-90 days.

Phase 3 (Months 7-9): The Ride-Along and RC&D Bet

The Phase 3 outcome target: close rate lifts 8-14 points (from 40-45% baseline to 50-58% across Comfort Advisors and senior techs), MPR — Membership Penetration Rate lifts from 22% baseline to 35-50%, recall percent drops from 5-7% industry median to under 3%, AI-powered RC&D triage classifies 100% of go-back tickets within 4 business hours with service-manager confirmation within 8 business hours.

The tool stack: Rilla for in-person sales coaching (the documented 18% close-rate lift via the 30-40 virtual ride-alongs per day per manager — Rilla's SiliconANGLE Dreamforce 2026 coverage benchmark); ResponsiBid for the AI quoting and good/better/best narrative drafting on $5K+ proposals; ServiceTitan / Sera / Housecall Pro native AI ticket classifiers for the RC&D (recall / callback / warranty) triage. Add to this: the AI root-cause clustering tool that produces the weekly Monday-morning heatmap by tech, part, symptom, and install date.

Phase 3 dependencies that gate launch: Tech Scorecard (Dimension 3) at 14+ of 20 — Rilla's scoring is wasted unless the service manager runs daily scorecard review and the advisor comp plan ties to close rate; Financing Baseline (Dimension 6) at 10+ of 15 — Rilla's close-rate lift partially depends on Wisetack / GreenSky / Synchrony soft-pull at the door producing approved-up-to amounts before the kitchen-table close. Below either gate, Phase 3 launches month 8-10 instead of month 7.

Phase 3 named owners: the service manager owns the daily Rilla transcript review (15-30 minutes daily) and the weekly coaching session with each advisor; the Comfort Advisor lead owns the close-rate-by-advisor scorecard and the comp-tied performance conversation; the office manager owns the AI RC&D classifier audit (weekly Monday 9 a.m. heatmap review confirms tagging accuracy); the owner reads the close-rate trend and the recall heatmap on Friday. Month 7 milestone: Rilla live, daily transcript review routine running, RC&D classifier active with manager confirmation. Month 8 milestone: close rate up 4-7 points, MPR at 30%+, recall trending to under 4%. Month 9 milestone: close rate up 8-14 points, MPR at 35-50%, recall at or under 3%, root-cause clustering producing actionable patterns per tech and per part.

Phase 3 risk register: advisor pushback on Rilla recording (covered in L4 Chapter 3 — the conversation, the comp plan, the pilot, the public win); two-party-consent state compliance on Rilla recordings (L4 Chapter 5 — disclosure scripts and consent forms required in 12 states by 2026); the close-rate lift is not free — it requires the service manager's daily 15-30 minute habit, which is the Phase 3 most-common failure mode (manager skips the review, scorecard sits unused, close rate flat).

Phase 4 (Months 10-12): The Attribution and Dashboard Bet

The Phase 4 outcome target: GLSA ROAS lifts 30-50% on top of the 3-4x baseline (so 4-6x ROAS achieved), cost-per-booked-call drops 15-25% by source, RPL — Revenue Per Lead — gets tracked weekly by source, Hatch reactivates 30-45% of stale leads producing 60-130 closed jobs in the quarter, the owner's 12-metric daily dashboard is live and refreshed daily. At a $4,400/week GLSA spend, the 30-50% AI bidding lift produces roughly $80K of annual incremental revenue.

The tool stack: CallRail Conversation Intelligence for lead-source attribution, missed-opportunity flagging, and sentiment alerting on every inbound call; Ryze AI (or comparable 2026 trade-specific GLSA bid platforms) for closed-revenue-tied bid optimization; NiceJob for review acquisition and customer-story engine; Podium AI Employee 2026 for review response, lead-capture chat, and SMS booking; Birdeye AI Employee 2026 for review response plus sentiment-monitoring early-warning; Yelp AI for Yelp-algorithm-optimized replies; Hatch for stale-lead reactivation with AI-drafted text and email sequences segmented by lead type. Add to this: the AI-assembled 12-metric daily dashboard pulling from ServiceTitan / Sera / HCP plus CallRail plus GLSA plus Hatch plus NiceJob.

Phase 4 dependencies that gate launch: Marketing Attribution Depth (Dimension 4) at 12+ of 15 — Ryze AI's closed-revenue feedback loop requires every inbound call routing through a tracked number with source attribution and ServiceTitan / HCP closed-revenue data feeding back to the bid engine; Owner Time and Routine Commitment (Dimension 5) at 10+ of 15 — the 12-metric dashboard is useless without the owner's daily 8-minute routine. Below the Dimension 4 gate, Phase 4 runs with partial scope: CallRail and NiceJob only, deferring Ryze AI bid management until attribution depth scores 12+ at the next quarterly audit.

Phase 4 named owners: the marketing manager owns the CallRail Conversation Intelligence weekly review (missed-opportunity flags, sentiment alerts, source attribution); the marketing manager owns the Hatch monthly report and segment design; the office manager or marketing coordinator owns the NiceJob / Podium / Birdeye review-response routine (60-second owner skim on every AI-drafted response); the owner owns the daily 12-metric dashboard read and the Friday recap. Month 10 milestone: CallRail full install, source attribution on 100% of inbound, Hatch live with first segment campaign. Month 11 milestone: Ryze AI live with closed-revenue feed, GLSA ROAS trending up 10-20%, Hatch first reactivation cohort at 25%+. Month 12 milestone: GLSA ROAS up 30-50%, owner's daily dashboard in 8-minute rotation, full 12-month roadmap retrospective delivered.

Phase 4 risk register: Ryze AI's feedback loop corrupts if negative-lead disputes are not run weekly (10-20% of GLSA spend leaks on non-service inquiries that the AI then learns are 'leads'); FTC endorsement-guideline exposure on AI-drafted review responses scales linearly with response volume — the 60-second owner skim is non-negotiable; Hatch reactivation can outrun CSR call-back capacity if not staged (the marketing manager and CSR floor must coordinate the sequence volume against capacity).

The Month-by-Month Milestone Table

The roadmap's defense artifact is the milestone table the owner walks into peer-group, coach, or PE-partner conversations with. One page, 12 rows (one per month), five columns: Month / Phase / Headline Metric / Supporting Metrics / Named Owner / Status.

Month 1 (Phase 1): Booking percent baseline, missed-call baseline, Avoca live, CSR daily review routine. Month 2 (Phase 1): Booking 72%+, after-hours 40%+, ServiceTitan summary AI live, CSR ACW at 30 sec. Month 3 (Phase 1): Booking 78-80%, missed-call under 10%, 90-day Avoca ROI delivered. Month 4 (Phase 2): Dispatch Pro live, override discipline 95%+, RPT baseline per truck. Month 5 (Phase 2): RPT up 6-10%, show rate 89%+. Month 6 (Phase 2): RPT up 12-18%, show rate 92%+. Month 7 (Phase 3): Rilla live, RC&D classifier active, daily transcript review routine. Month 8 (Phase 3): Close rate up 4-7 pts, MPR 30%+, recall under 4%. Month 9 (Phase 3): Close rate up 8-14 pts, MPR 35-50%, recall under 3%. Month 10 (Phase 4): CallRail full install, Hatch live, source attribution 100%. Month 11 (Phase 4): Ryze AI live, GLSA ROAS up 10-20%, Hatch first cohort 25%+. Month 12 (Phase 4): GLSA ROAS up 30-50%, owner dashboard live, 12-month retrospective.

The table is the most-frequently-referenced artifact in the L4 capstone defense. Every claim in the AI roadmap ties to a row in the table. Every milestone misses get explained with a documented dependency that did not close (FSM data hygiene slipped, service manager left in month 6, Dimension 4 score did not move in Q2 re-audit). The artifact converts AI strategy from narrative to evidence.

Quarterly Re-Audit and Roadmap Correction

The roadmap is not static. The readiness audit re-runs every 90 days at each phase boundary. The owner reviews the composite movement against the prior quarter and the phase-launch decisions against the new score. Phase 2 may launch on schedule, on a 30-day delay, or with partial scope based on Q2 re-audit. Same logic for Phase 3 at Q3 re-audit and Phase 4 at Q4 re-audit.

Quarterly correction examples. A shop launches Phase 1 on schedule at Band B; Q2 re-audit shows Dimension 3 (Tech Scorecard) only moved from 11 to 13 instead of the projected 15 — Phase 2 launches on schedule but with a tighter service-manager training overlay in month 4. A shop launches Phase 1 month 4 at Band C; Q3 re-audit shows Dimension 6 (Financing) still at 8 despite intent to fix — Phase 3's Rilla launch defers to month 10 and the financing-process remediation gets named as the Q3 priority. A shop launches Phase 4 on schedule but Q4 re-audit shows Marketing Attribution at 10/15 — Ryze AI partial-scope launch with CallRail-only feedback loop until next quarter when attribution lifts.

Quarterly correction is the discipline that keeps the roadmap honest. The owner who reports the quarterly composite, the dependency gaps, and the corrected timeline to the coach, peer group, or PE partner gets credibility. The owner who reports "all phases on schedule, all metrics moving" without supporting audit movement gets skepticism. Quarterly correction is the operating-cadence rigor that distinguishes top-quartile AI execution from median or below.

The 12-Month Defense of the Roadmap

By month 12, the owner has the artifact stack for the L4 capstone defense and the platform reporting. Readiness audit composite trajectory: Band C start (52) to Band A end (82). Four phase ROIs documented: Phase 1 produced $280K of recovered revenue at $35K of tool and process cost (8x payback); Phase 2 produced $420K of incremental RPT at $50K cost (8.4x); Phase 3 produced $310K of close-rate lift plus $60K of recall reduction at $45K cost (8.2x); Phase 4 produced $95K of GLSA incremental at $40K cost (2.4x — long-tail compounding). Total year-one: $1.17M of attributed revenue lift at $170K of AI tool and process spend (6.9x).

The owner walks into the Nexstar peer call with the audit trajectory, the phase ROIs, the milestone table, and the next-12-month plan. The conversation moves from 'are you doing AI?' to 'how is your readiness composite trending and what is the Q2 priority?' For platform-owned shops or franchisees reporting to Wrench Group, Authority Brands, Apex Service Partners, Sila Services, Path Light Pro, or Redwood Services HQ, the same artifact stack feeds into the platform QBR slide deck with EBITDA waterfall by location and AI-ROI line items by tool. For PE partners, the same stack supports the quarterly synergy synthesis and the board's KPI-by-location review.

The 12-month roadmap is not a sales document. It is an operating substrate. The shops that execute it at Band A by year-end have a different P&L 24 months out than shops that bought tools without a roadmap. The 12-month plan is what makes the AI bet defensible to the team, the customer, the coach, the peer group, the franchise HQ, and the PE partner. The audit produces the score. The roadmap produces the plan. The quarterly correction keeps the plan honest. The 12-month retrospective produces the artifact for year two.

Key Takeaways

• The 12-month roadmap divides into four 90-day phases. 90 days aligns to AI ROI visibility, change-management absorption capacity, and existing quarterly reporting cadence (Nexstar, BDR, franchise HQ, PE board).
• Phase 1 (months 1-3): Avoca / Jobber AI Receptionist / Housecall Pro AI Agents / ServiceTitan Voice plus call-summary AI. Booking percent lifts from 65% to 80%, missed-call drops below 8%, after-hours capture above 60%. Gates on FSM Data Quality (12+/20) and CSR Maturity (9+/15).
• Phase 2 (months 4-6): Dispatch Pro / Sera / FieldEdge plus dispatch override discipline plus service-manager scorecard cadence. RPT lifts 12-18%, dispatch yield improves, show rate moves to 92%+. Gates on Tech Scorecard (12+/20) and FSM Data (14+/20).
• Phase 3 (months 7-9): Rilla, ResponsiBid, AI-powered RC&D triage. Close rate lifts 8-14 points, MPR moves to 35-50%, recall percent drops below 3%. Gates on Tech Scorecard (14+/20) and Financing Baseline (10+/15).
• Phase 4 (months 10-12): CallRail Conversation Intelligence, Ryze AI for GLSA bidding, NiceJob, Podium AI Employee, Birdeye AI Employee, Hatch, Yelp AI, plus the owner's 12-metric daily dashboard. GLSA ROAS lifts 30-50% on top of 3-4x baseline. Gates on Marketing Attribution (12+/15) and Owner Routine (10+/15).
• Each phase produces one headline metric, two supporting lifts, one daily artifact. Phase 1: AI-summarized booking log. Phase 2: daily dispatch-override log. Phase 3: weekly ride-along scorecard with AI-clustered coaching topics. Phase 4: owner's 12-metric daily dashboard.
• The roadmap is gated by the quarterly readiness audit. Each phase boundary is a re-audit checkpoint. Below-gate dimensions delay or partial-scope the next phase. Above-gate dimensions launch on schedule.
• Year-one ROI on a successful execution: approximately $1.17M attributed revenue lift at $170K of AI tool and process spend (~6.9x payback). Phase 1: ~8x. Phase 2: ~8.4x. Phase 3: ~8.2x. Phase 4: ~2.4x (long-tail compounding).
• The roadmap defends in front of any reviewer. Nexstar peer calls, CertainPath, BDR, franchise HQ QBRs (Wrench Group, Authority Brands, Apex Service Partners, Sila Services, Path Light Pro, Redwood Services), and PE diligence. The milestone table is the most-referenced artifact in the defense.