By 2028, voice agents at the trades will not just answer the phone. They will quote, schedule, run the financing soft-pull, pitch the membership, and close the dispatch fee — all before a human CSR sees the ticket. Avoca, Rilla, Hatch, and ServiceTitan are each building toward the same convergence point from different starting positions: Avoca from inbound voice, Rilla from in-home advisor recording, Hatch from outbound nurture, ServiceTitan from the platform-system-of-record. The four vendors will not produce four products in 2028 — they will produce roughly the same product under four brand names with four go-to-market motions. For the platform CEO running 25, 100, or 450 locations, that convergence is the most important AI prediction of the second half of the decade. It rewrites the CSR role at 80% AI inbound. It collapses the dispatcher role to exception handling on a board that re-solves every 90 seconds. It moves the human-AI handoff from "the AI books, the human runs the call" to "the AI runs the call, the human inherits an exception." This lesson is the L5 strategic read on the convergence and how the platform owner positions for it now — capital plan, org chart, vendor negotiation, operating-bench day-to-day. The platform CEO who reads this and adjusts the 2027 plan walks into the PE board with a 24-month positioning thesis rather than a 12-month tactical update.

The Four-Vendor Convergence Thesis

The 2026 voice-AI tier is four named vendors with four product positions. Avoca runs inbound voice — answer rate, missed-call recovery, after-hours capture, basic warm transfer to a human CSR for everything beyond booking. The HL Bowman case study (100% answer rate, 70% YoY revenue growth, cost per conversion $350 to $215, 39% reduction) is the public proof point. Avoca's April 2026 $125M Series B at $1B valuation funds the product roadmap explicitly toward more transactional capability. Rilla runs in-person and virtual ride-along recording — sales-coaching transcript analysis, advisor scorecard generation, close-rate lift through coaching loops. The Climate Experts 18% close-rate lift and the 30-40 virtual ride-alongs per manager per day throughput are the proof. Rilla owns the audio-of-the-conversation tier. Hatch runs outbound nurture — stale-lead reactivation, dormant-pile re-engagement, multi-touch SMS and voice cadences with 30-45% reactivation lifts. ServiceTitan runs the platform — Titan Intelligence, Dispatch Pro, Voice, the system-of-record for the conversation, the ticket, the price book, the payment, the warranty registration.

The four positions look different in 2026. By 2028, each vendor builds toward the same product surface: an AI voice agent that handles the full inbound-to-booked-job conversation — quote, scheduling, financing soft-pull, membership pitch, dispatch fee close — autonomously, with human CSR inheriting only exceptions. Avoca arrives from answer-rate and adds transactional capability. Rilla arrives from conversation-analysis and extends to real-time inbound. Hatch arrives from outbound-cadence and adds inbound. ServiceTitan arrives from platform-of-record with native voice-tier integration. The four overlapping surfaces are no longer differentiated by what they do — they are differentiated by which platform data they sit closest to and which vendor-management muscle the CEO has invested in.

The convergence is forecast against vendor capital, talent, competitive pressure, and go-to-market math. Avoca cannot defend a $1B valuation in 2028 on inbound-answer alone — the TAM is bounded by residential trades inbound-call volume, and the lift from 22% missed-call baseline to under 5% is a one-time gain. To justify the next funding round, Avoca's product must capture more of the booked-call conversation. Rilla cannot grow the ride-along TAM beyond Comfort Advisor seat count; to grow into the larger inbound-conversation TAM, Rilla extends to live calls. Hatch's nurture TAM is bounded by stale-lead pile volume; the path is to capture the live inbound Avoca owns. ServiceTitan is already at the platform level; the path is to make Titan Intelligence's voice tier as good as Avoca's so customers stop buying both. Four vendors, four growth pressures, one convergence destination.

What the 2028 Voice Agent Actually Does

The 2028 voice agent picks up the no-heat call at 6:47 a.m. It identifies the caller from the inbound number against the platform CRM (ServiceTitan, Sera, HCP). It knows the equipment from prior tickets, membership status, prior dispatch-fee waivers, financing history. The agent asks diagnostic questions in the platform's brand voice — equipment age, symptom, urgency, household occupancy (small kids, elderly, oxygen dependence, no-heat-with-newborn priority routing). It quotes the dispatch fee at the platform's flat rate, waives it for the membership tier the customer qualifies for, and books the appointment in a Dispatch Pro-optimized slot.

It runs the financing soft-pull if the conversation indicates a likely high-ticket diagnostic (system over 12 years, recurring failures, customer-stated upgrade interest). Wisetack, GreenSky, or Synchrony produces an approval tier in 60 seconds; the agent quotes the payment range conversationally. It pitches the membership pivot for non-members on aged equipment ("our $39-a-month membership covers your diagnostic today and discounts any repair or replacement"). It drafts the on-my-way text with the tech's name, photo, ETA, and review-prompt link before the call ends. It logs the structured call summary, customer-mood read, upsell flags, and exception triggers to the platform CRM.

Handoff to the human CSR happens for exceptions: agent confidence on an answer below 0.85, customer dissatisfaction or escalation cues, customer requests a human (the platform supports this on demand), or call scope exceeds the agent's defined surface (multi-system commercial, insurance claims, code questions, dispute resolution). Target exception rate at the high end of platform performance: 15-20% by 2028 — meaning 80-85% of inbound clears the agent without human CSR involvement. The CSR floor is repositioned, not eliminated. The CSR becomes exception handler, senior-call escalation point, dispute resolver, high-judgment conversation operator. Comp plans adjust accordingly.

What the CSR Floor Looks Like at 80% AI Inbound

The platform CSR floor in 2026 averages 8-15 CSRs per 100 trucks. The CSR's day is roughly 60-90 calls handled, 65% booking rate, 12% missed-call leakage in the median shop. The CSR's KPIs are booking %, average revenue per booked call, after-call work time, customer satisfaction score, and adherence to the script. The role is high-volume, high-script, moderate-judgment. Comp ranges $18-$28/hour with bonus structures tied to booking % and revenue per call.

By 2028 at 80% AI inbound, the platform CSR floor shrinks in headcount and changes in skill profile. A 100-truck platform that ran 12 CSRs in 2026 runs 4-5 in 2028. The remaining CSRs handle the 15-20% exception volume — about 12-18 exception calls per CSR per day, each higher-judgment than the average 2026 call. KPIs shift: exception-resolution rate, customer-satisfaction on escalated calls, AI-handoff audit quality, upsell capture on senior-call escalation, and AI-feedback-loop contribution (CSR notes on AI failure modes that inform vendor product tuning). Comp ranges shift to $25-$40/hour with higher base and lower variable.

CSR-floor governance also shifts. Centralized shared-services floors across multi-brand portfolios become more economical because exception volume aggregates — a single 20-CSR shared services center serves a 300-truck platform that would have required 35-45 distributed CSRs in 2026. The Director of AI Operations governs agent tuning, exception-routing logic, and AI-handoff quality discipline. The shared services floor reports to a CSR Operations Lead more senior than the 2026 floor manager, governing exception SLAs (escalation response time, handoff audit cadence, dispute resolution time). The CSR career path tilts from booking-rate volume to senior-judgment exception handling, with progression into Conversation QA Lead or Prompt Librarian roles.

The transition is a 2026-2028 ramp, not a step-function. Platforms running the four-stage rollout framework (L5 Ch1 Lesson 1) move CSR floors through the transition stage by stage — comp-plan rewriting, career-path documentation, hiring-profile adjustment, training redirection, and severance budget for the CSRs whose roles disappear without transitioning. The platform CEO who plans the restructuring across 2026-2028 produces a smaller, more skilled, better-compensated CSR floor that the AI agent depends on. The CEO who doesn't arrives at 2028 with a floor too large, a comp plan that incentivizes obsolete metrics, and a turnover spike when the floor realizes the AI is taking the easy calls.

What the Dispatcher Role Looks Like When the Board Re-Solves Every 90 Seconds

The 2026 dispatcher works the board manually with Dispatch Pro running 10-minute re-evaluation cycles in the background. The dispatcher's day is roughly 200-400 tile movements, 30-60 override decisions, 5-10 escalations to the GM or operations lead, and continuous radio coordination with the field. The dispatcher's KPIs are revenue per truck per day (RPT), dispatch yield, on-time arrival rate, callback recovery, and override accuracy (when an override produced a better outcome than the AI suggestion). The role is high-tempo, high-judgment, very-low-glamour, deeply skilled — the dispatcher is the single most underpaid role in most shops relative to their EBITDA impact.

By 2028, Dispatch Pro and the equivalent products (Sera, FieldEdge AI, ServiceTitan Dispatch Pro 2028) re-solve the board every 60-90 seconds rather than every 10 minutes. The frequency change is not a small feature — it changes the role fundamentally. The board reflects current state at all times; the dispatcher does not work the board, the dispatcher governs the board. The dispatcher's day shifts from 200-400 tile movements to 20-50 override-and-exception decisions, each higher-judgment. KPIs shift to override quality (overrides that produced better outcomes than the AI's optimized board), exception-resolution time, AI-feedback contribution, and revenue per truck across the portfolio.

Dispatcher headcount changes less than CSR headcount. A 100-truck platform that ran 4-6 dispatchers across shifts in 2026 runs 3-5 in 2028 — exception complexity grows even as routine workload shrinks. Pay grows. The senior dispatcher is the most experienced operational judgment in field operations after the GM; the role becomes Dispatch Operations Lead or Field Operations Director with comp in the $90K-$140K range plus bonus on RPT and exception metrics. Titles may rename to Field Operations Analyst, AI Dispatch Operator, or Dispatch Quality Lead.

The board's re-solve frequency also shifts how the platform measures dispatch performance. RPT becomes a real-time metric rather than an end-of-day metric — the dispatcher's screen shows current RPT against target throughout the day, with AI flagging trajectory deviations early. Dispatch yield aggregates across the portfolio's 100+ trucks in near-real-time; the quarterly board review includes dispatch yield by region, brand, and truck type. The Director of AI Operations governs Dispatch Pro tuning; the dispatcher operations lead governs override discipline; the CEO reports dispatch contribution to EBITDA on slide 4 of the quarterly deck. Dispatch becomes a margin lever, not a back-office function.

How the Platform CEO Positions the Capital Plan for the Convergence

The 2027-2028 capital plan for a platform CEO has three voice-agent positioning lines that didn't exist in the 2024-2025 plan. First, AI voice contract evolution — the 2026 Avoca contract is priced per booked call or per location seat in the $1K-$3K/month range. The 2028 voice agent will be priced per closed transaction or per share of the conversation flow at materially higher per-unit cost. The platform CEO budgets for the contract evolution and negotiates the multi-year contract structure that protects against unbounded price growth (deployment-success-tied pricing, volume-tier discounts, multi-year rate locks).

Second, CSR-floor restructuring cost. The transition from 12 CSRs to 4-5 across a 100-truck platform produces severance, retraining, and recruiting costs. The platform budgets for restructuring as a multi-quarter expense running alongside agent deployment. The CFO and Director of AI Operations co-build the transition plan with named metric checkpoints (exception rate below 25% triggers next floor reduction; customer-satisfaction above benchmark triggers next capability expansion; handoff quality above 0.90 triggers next shared-services consolidation).

Third, vendor portfolio rebalance. The 2026 platform may run Avoca + Rilla + Hatch + ServiceTitan Voice in parallel because surfaces don't overlap. By 2028 the surfaces overlap substantially. The CEO decides whether to run two vendors (risk hedge, negotiating leverage), one vendor (integration simplicity, deeper discount), or a primary-secondary structure (Avoca primary, ServiceTitan Voice as the integration-native backup, Rilla retained for advisor coaching and the recording-transcript layer). The 2028 vendor portfolio is materially different from 2026; the capital plan funds the transition. Decision frame: where does the AI moat strengthen most — vendor-dependent (Avoca product depth) or platform-dependent (ServiceTitan-integrated, multiple voice-tier options)? Platform CEO answers explicitly in the 2027 board plan.

The capital plan also funds operator-bench investment. The Director of AI Operations role expands; the Conversation QA Lead role emerges (governs AI agent quality); the Prompt Librarian manages prompt versioning across brands; the Data Center Account Director (where the platform has electrical capacity) builds the 2027-2030 hyperscale positioning thesis. The 2027-2028 platform org chart has 3-6 new senior roles that did not exist in 2025; the CEO recruits ahead of capability deployment. The framework's third-order benefit makes the bench investment compound across multiple workflows.

Vendor Negotiation Posture in the Convergence Era

The 2026 vendor negotiation posture is bilateral — platform and vendor negotiate a contract for a defined product surface. The 2028 vendor negotiation posture is portfolio-level, multi-vendor, and convergence-aware. The platform CEO walks into the 2028 Avoca renewal conversation with three pieces of leverage that didn't exist in 2026. First, the platform now has stage-gate-measured data on Avoca's deployment success at scale across 80+ locations; the platform is no longer a customer reference, the platform is a co-roadmap partner. Second, the platform has alternative vendor proof — ServiceTitan Voice's 2027-2028 capability set, Rilla's live-call extension, Hatch's inbound capability — meaning Avoca is no longer the only voice-agent option. Third, the platform has operator-bench depth that can absorb a vendor switch; the switching cost is lower in 2028 than in 2026 because the deployment artifacts (CSR training, GM onboarding, handoff protocols) translate across vendors.

The 2028 vendor contract structure includes deployment-success-tied pricing, multi-year rate locks with volume tier discounts, vendor-funded customer success investment, contractual product-roadmap commitments tied to platform-level deployment learnings, and exit/migration clauses that protect against vendor capture. The platform CEO who negotiates this structure compounds leverage quarterly. The platform CEO who treats each vendor renewal as a transactional reprice loses leverage and absorbs vendor price growth.

The convergence also forces a vendor-portfolio decision the platform CEO presents to the PE board. The decision is not "which voice-agent vendor wins?" The decision is "which vendor-portfolio strategy fits our platform's compound advantage?" Three coherent strategies emerge by 2028. Strategy A: single-vendor depth (Avoca as the primary voice agent, ServiceTitan as platform-of-record, no second voice vendor). Lower integration overhead, deeper discount, vendor-capture risk. Strategy B: dual-vendor hedge (Avoca + ServiceTitan Voice, with workload split by call type or by region). Higher integration overhead, lower vendor-capture risk, ongoing optionality. Strategy C: platform-native (ServiceTitan Voice as primary, Avoca retained for advanced features, Rilla retained for the recording-transcript layer). Maximum integration leverage, vendor-portfolio rebalancing toward the platform owner. Each strategy is defensible; the platform CEO picks one explicitly and defends it on slide 6 of the 2028 PE board deck.

How the Platform Owner Positions the 2027-2028 Thesis for the Board

The PE board reads the voice-agent convergence as either a thesis the platform CEO has anticipated or as a tactical update the platform CEO is reacting to. The framing matters. The thesis-version of the conversation: "The four voice-agent vendors converge on a single product surface by 2028. We have positioned the platform to capture the convergence by deploying Avoca across 80 of 100 locations in 2026-2027 under the stage-gate framework, by building the operator bench that can absorb vendor portfolio changes, by negotiating multi-year contracts with deployment-success-tied pricing, and by sequencing the CSR-floor restructuring across 2027-2028 with named metric checkpoints. The 2028 EBITDA contribution from voice-agent automation is $X.X million across the portfolio, vs. $X.X million in 2026. The 2028 CSR-floor headcount is Y vs. Z in 2026 — a $X.X million annualized labor cost reduction that funds the AI vendor contract and the operator-bench investment with margin to spare. We are positioned to lead the convergence rather than absorb it."

The tactical version: "Avoca's product is changing and we're evaluating renewal; CSR turnover is elevated; we're not sure whether to add ServiceTitan Voice; the bench is stretched." Same operational reality, opposite framings. The PE board reads the first as operational maturity worth multiple expansion at exit; the second as a roll-up team that didn't see the convergence coming. Multiple expansion at exit is the single largest financial outcome the CEO drives; the framing of the AI thesis is one of the highest-leverage inputs into that exit multiple.

The thesis preparation starts in 2026. The L5 capstone plan (Ch7 Lesson 1) is where the CEO commits the 90-day plan that initiates it. Voice-agent convergence is the single largest 2027-2028 operating reality the platform CEO will navigate. The framework, operator bench, vendor portfolio, and CSR-floor restructuring all compound toward the same outcome: a platform that runs convergence as an operational compounding asset rather than a tactical scramble.

Key Takeaways

• Four vendors converge on one product surface by 2028: Avoca (inbound voice → adds transactional), Rilla (ride-along audio → extends to live calls), Hatch (outbound nurture → adds inbound), ServiceTitan (platform-of-record → matches Avoca depth). Four growth pressures, one convergence destination — the AI voice agent that handles inbound conversation end-to-end through booking, soft-pull, membership pivot, and dispatch close.
• The 2028 voice agent runs the full call: identifies caller from CRM, asks diagnostic questions, quotes dispatch fee, books in Dispatch Pro-optimized slot, runs financing soft-pull, pitches membership pivot, drafts on-my-way text, logs structured summary — autonomously, 80-85% of inbound, with human CSR inheriting only the 15-20% exception tier.
• CSR floor restructures from 12 per 100 trucks to 4-5 at 80% AI inbound. Comp shifts $18-$28/hour to $25-$40/hour (higher base, lower variable, exception-handling premium). KPIs shift from booking % to exception-resolution rate, AI-handoff audit quality, and senior-call upsell capture. Centralized shared-services CSR floors aggregate exception volume across multi-brand portfolios.
• Dispatcher role compresses to exception handling and override quality as Dispatch Pro re-solves the board every 60-90 seconds rather than every 10 minutes. Headcount drops modestly (4-6 to 3-5 per 100 trucks); pay rises ($90K-$140K with bonus on RPT and exception metrics). The dispatcher becomes a Field Operations Lead or AI Dispatch Operator.
• Three capital-plan lines for the 2027-2028 convergence transition: AI voice contract evolution (deployment-success-tied pricing, multi-year locks), CSR-floor restructuring cost (severance, retraining, recruiting across multiple quarters), and vendor portfolio rebalance (one-vendor depth, dual-vendor hedge, or platform-native primary).
• Vendor negotiation posture shifts from bilateral to portfolio-level by 2028. Platform CEO has stage-gate-measured deployment data, alternative-vendor proof, and operator-bench-supported switching capacity. Contract structures include deployment-success-tied pricing, multi-year rate locks with volume tiers, vendor-funded customer success investment, and exit/migration clauses.
• Three coherent vendor-portfolio strategies by 2028: Strategy A single-vendor depth (Avoca + ServiceTitan, no second voice vendor — lower overhead, higher vendor-capture risk). Strategy B dual-vendor hedge (Avoca + ServiceTitan Voice split by call type or region). Strategy C platform-native (ServiceTitan Voice primary, Avoca retained for advanced features, Rilla for recording-transcript layer). Pick one explicitly on slide 6 of the 2028 board deck.
• Operator bench investment is the compound advantage: Director of AI Operations expands, Conversation QA Lead emerges, Prompt Librarian governs versioning across brands, Data Center Account Director builds the 2027-2030 hyperscale thesis. The 2027-2028 org chart has 3-6 new senior roles that did not exist in 2025; the framework's third-order benefit (L5 Ch1 Lesson 1) makes the bench investment compound across workflows.
• The PE board reads the convergence as thesis or tactical. Thesis framing produces multiple expansion at exit; tactical framing produces roll-up multiples. The platform CEO who walks into the 2028 board review with a 24-month positioning thesis captures both the convergence margin and the multiple. Tactical framing concedes both.
• The convergence is not 2028 step-function — it is a 2026-2028 ramp. Platforms running the four-stage framework move CSR floors, dispatcher roles, vendor portfolios, and capital plans through the transition stage by stage. The platform CEO who plans the ramp in 2026 captures the convergence; the platform CEO who doesn't gets caught flat-footed when Avoca's 2028 product crosses the quote-and-close line and resets industry CSR-floor expectations overnight.