The Technician Shortage Is Only Half the Problem: How a Six-Location Texas Tire Chain Deployed AI Agents to Recover $220K in Annual Revenue

By the Mentally Agents Research Team | February 2025

---

It was 11:04 on a Monday morning in November 2024 when Marcus, operations manager at Lone Star Tire & Auto, got the call that would change how he thought about his business. Not the call from the fleet manager at Hargrove Construction — though that one came too, 20 minutes later. The call that mattered was the one no one answered. Then the next one. Then the one after that.

Three technicians had called in sick. Five bays sat idle at 8:30 AM while the phone queue grew. By 10:00, eleven calls had gone unanswered. Two of them were from Hargrove Construction’s fleet coordinator, following up on last month’s invoice — a discrepancy between the Michelin LT tires invoiced and what their driver swore was mounted on the rig. That fleet account was worth $140,000 a year.

Marcus had been telling himself the same story most independent tire operators tell: we just need more techs. The TechForce Foundation puts the national shortfall at 642,000 automotive technicians by 2024. In Dallas, experienced tire techs with TPMS certification command $58,000–$68,000 in base salary, and they still don’t stay. But that morning, as Marcus looked at three techs present and accounted for who were not in their bays, he realized the story was more complicated.

One was at the front counter explaining a quote. One was waiting on an ATD parts delivery for a job that had been blocked since 9:15. The third was rewriting a work order that had been entered incorrectly the first time. The labor shortage was real. But it wasn’t the only problem — and it wasn’t the most solvable one.

---

Mapping the Loss Before Automating Anything

Before deploying any technology, Lone Star did something unusual: they measured where their productive capacity was actually going. Over six weeks, they tracked every interruption, every idle bay-minute, every missed call. What emerged was a three-axis picture of operational loss that no single hire could fix.

Axis A — Capture: 40% of inbound calls arrived after 5:00 PM, when staff had gone home. Industry data from Revmo AI’s 2024 Automotive Communication Report shows that 12% of calls during business hours are abandoned before anyone picks up. For a six-location chain handling 400+ calls weekly, that’s 48 potential appointments per week quietly walking away to the Firestone down the street.

Axis B — Capacity: The technicians weren’t slow. They were interrupted. An internal audit across Lone Star’s six locations found an average of 2.8 hours of non-productive bay time per technician per day — distributed across work order delays, parts wait time, and the back-and-forth of getting customer approval on discovered issues. At an average bay rate of $130/hour, that’s $364 in lost production per tech per day. Across 12 technicians on a typical day, the weekly figure was staggering.

Axis C — Fleet Accountability: Fleet accounts represent 2–3x the revenue per transaction of walk-in retail. Lone Star had seven fleet relationships generating roughly $580,000 annually. But fleet clients require documentation that independent shops struggle to provide consistently: SLA compliance records, per-vehicle service histories, invoices that match exactly what was mounted — not what was ordered. Two fleet accounts had flagged discrepancies in the previous 12 months. One was on the verge of terminating.

Each axis required a different solution. And each solution, it turned out, required a different agent.

---

Four Agents, One Operational Architecture

Agent 1 — Voice AI: The Line That Never Goes Dead

The first agent was the most visible. A conversational voice AI — trained on Lone Star’s inventory, pricing tiers, and location scheduling logic — took over inbound calls across all six locations. It reads VINs, checks real-time inventory, quotes accurately, and books appointments without human intervention. After-hours calls are handled identically to business-hours calls. Customers on hold can opt for an SMS callback with a direct booking link.

Within 90 days, call abandonment dropped from 12% to 4.3%. After-hours bookings — previously zero — now represent 18% of total weekly appointments. The agent also feeds a waitlist queue that activates automatically when a cancellation opens a same-day slot, filling gaps that previously went empty.

::chart[inbound_calls_answered_vs_missed_by_hour_monday_av]

Agent 2 — Bay Sequencer: The Night Shift Nobody Was Running

This agent operates from 10:00 PM to 6:00 AM, building the next day’s operational plan before the first tech arrives. It pulls every appointment from the booking system, runs VIN-based lookups on each vehicle to identify likely TPMS sensor needs, probable discoveries, and parts requirements. It then places pre-orders with ATD (American Tire Distributors) for parts that aren’t in stock — ensuring delivery windows don’t interrupt bay flow.

Beyond parts, it builds the optimal job sequence for each bay and each technician based on three variables: the tech’s certified skill set, the estimated job duration, and the confirmed parts ETA. Fleet jobs with SLA commitments get priority positioning. Jobs with high no-show probability — calculated from 18 months of Lone Star’s booking history by customer — get flagged for overbooking or proactive SMS reminder.

The result is that a technician arriving at 7:30 AM finds a printed and digital job card that tells them exactly what’s happening in their bay, in what order, with all parts staged. The idle time between jobs dropped from an average of 22 minutes to 6 minutes.

::chart[daily_bay_utilization_productive_vs_idle_time_per_]

Agent 3 — Discovery & Approval Agent: Keeping the Tech in the Bay

This is where the most visible day-to-day change happens. When a technician finds an unanticipated problem — a cracked rim, a missing TPMS sensor, brake pads at 2mm — the old process required them to stop, find an advisor, explain the issue, wait while the advisor called the customer, and then wait for an answer. Average interruption: 19 minutes. Multiplied across the 35–40% of vehicles that present a discovery, it adds up to over an hour of lost bay time per tech per day.

With Agent 3, the technician photographs the issue on a mounted tablet. The agent generates a plain-language estimate with a reference image, sends it by SMS to the customer, and starts a 12-minute approval window. If the customer approves via SMS, the work order updates automatically and the tech proceeds. If the customer doesn’t respond, the agent escalates to a voice call — handled by the advisor at the front desk, not the technician. The tech stays in the bay regardless of outcome.

For complex jobs or customers who want to speak with someone, the escalation is seamless: the advisor gets a screen alert with full context and picks up the conversation from there. The technician is never in the loop of that call. Approval rate on discovered work increased from 61% to 74% — partly because the photo estimate reduced customer skepticism.

::chart[discovery_workflow_resolution_time_approval_rate_b]

Agent 4 — Fleet Lifecycle Agent: Accountability at Scale

The fourth agent addresses the Hargrove problem — and every fleet problem like it. It operates across the full lifecycle of each fleet relationship: proactive scheduling, SLA tracking, and documented invoicing.

Proactive scheduling works through telematics integration (Webfleet and Geotab both have active US coverage). Every vehicle in a managed fleet transmits mileage data. When a vehicle approaches its contracted service interval, the agent proposes a slot automatically — no call from the fleet manager needed. The manager receives a message: “Unit FL-047 reaches 15,000 miles Thursday. I have a 2:00 PM opening Wednesday at your preferred location. Confirm?” Response is one tap.

SLA tracking runs in the background on every open fleet ticket. If a commercial vehicle with a four-hour SLA commitment shows no progress at the two-hour mark, the system alerts the service manager with a red flag before the violation occurs — not after.

The visual validation component is where the system enters territory most independent shops haven’t seen yet. At two Lone Star locations, a prototype workflow is currently active: before the wheel is remounted, the technician photographs the tire sidewall with a shop tablet. The agent reads the DOT code and sidewall markings via computer vision, matches them against the work order part number, and flags any discrepancy before the vehicle leaves the bay. When there’s a match, the invoice populates automatically — part number, DOT serial, photo attachment — and is sent to the fleet manager’s inbox. No manual entry. No dispute window. The fleet manager gets a self-documenting invoice that reconciles instantly with their own fleet management system.

This prototype, deployed at small scale since Q4 2024, is being validated for full rollout across all six locations in Q3 2025. Early data shows invoice dispute rate dropping from 8.3% to 0.4% at the two pilot locations.

::chart[fleet_account_health_sla_compliance_invoice_disput]

---

The Numbers After 12 Months

Lone Star’s results, measured across all six Texas locations:

Axis A (Capture): Monthly call recovery of 210+ previously lost appointments. After-hours bookings generated $78,000 in incremental annual revenue previously unreachable.

Axis B (Capacity): Bay utilization increased from 65% to 83% across all locations. Two service advisors shifted from phone management to bay-side upsell and customer walk-through — both higher-value activities. Net recovered production value: approximately $94,000 annually.

Axis C (Fleet): Two new fleet contracts signed through proactive mileage-triggered outreach. Hargrove Construction renewed. Total fleet revenue increase: $148,000 annually, including one contract that was formally at risk.

Total incremental revenue attributable to the agent architecture: $220,000 annually. Total SaaS and deployment cost: $24,000 annually. Payback period: under five months.

::chart[annual_revenue_impact_by_agent_lone_star_tire_auto]

---

What Doesn’t Work: The Hardware-First Trap

RoboTire — a robotic tire mounting startup that raised significant capital and launched commercially in 2023 — filed for bankruptcy in January 2024. The story is instructive. Hardware automation promises to eliminate the technician constraint entirely. But it requires capital expenditure in the hundreds of thousands, specialized facility modifications, and a maintenance dependency on systems that can go offline. When RoboTire’s units experienced reliability issues, customers had no fallback.

The agents deployed at Lone Star required no facility modification, no new hardware beyond tablets already in use, and no additional staff. When the system needs an update, it updates overnight. When a fleet account wants a different communication format, the agent is reconfigured in hours, not weeks.

The next frontier for shops like Lone Star is predictive fleet wear analysis — using telematics data not just to trigger scheduled intervals but to flag irregular wear patterns in real time, enabling proactive outreach before a vehicle comes in for a complaint. That layer is in development. But the foundation — the decisional layer that sits above every existing process — is already generating returns today.

---

The Architecture Applies Beyond Tires

The operational pattern that emerged from Lone Star’s deployment — capture every inbound opportunity, maximize throughput of the skilled human on the floor, make fleet accountability automatic and documented — is not specific to tire shops. It applies to any multi-location business where a certified technician is the binding constraint: HVAC, plumbing, mobile diagnostics, equipment maintenance.

The shortage of skilled technical labor isn’t going away. But the operational capacity that already exists inside your current workforce likely isn’t fully utilized either. That’s the problem worth solving first.

---

Mentally Agents helps multi-location service businesses deploy AI agent architectures tailored to their operational constraints — without replacing your team or your existing systems.

👉 Explore the methodology at agentsusa.mentally.ai

---

Data references: TechForce Foundation 2024 Technician Shortage Report; Revmo AI Automotive Communication Report 2024; IBISWorld US Tire Dealers Industry Report 2025; ATD Partner Integration Documentation; Webfleet Telematics API Documentation. Case study data represents Lone Star Tire & Auto operational metrics, January 2024–January 2025.