Maximizing CP Gross & Shop Throughput
Without Adding Headcount.
AI solutions for diagnosed revenue leakage.
01The Bay Bottleneck
02The Advisor Bottleneck
03The Aftermarket Bleed
Demo Mandar Gokhale April 2026
Speaker notes
  • Read the headline, don't pitch it. Paul pattern-matches CP Gross + headcount instantly.
  • One-line tee-up: "three places diagnosed revenue leaks today — we've built an AI product for each." Name them. Don't explain them; Paul already knows.
  • Hand-raise test: "these the three you'd name?" — if he says yes, you've calibrated. If he names a fourth, listen.
  • Move to slide 2 within 60 seconds.
Recite · Technician walkaround
BOTTLENECK 01 · THE BAY
Gloves on. Phone on the toolbox. Inspect by voice.
HOW
  • Plate scan + registry lookup identifies the vehicle — right inspection profile auto-loaded for the make + model
  • Tech speaks mechanic shorthand while working — AI transcribes + classifies on-device
  • Every point populated, severity inferred
  • Auto-pushes the MPI into OEM compliance portals
WHY IT WINS
  • 10–15 minutes faster per RO — hours of billed capacity unlocked weekly
  • Pencil-whipping ends — no dropdowns to click with gloves off
  • Warranty Admin stops double-entering into OEM portals
Speaker notes
  • Start here because it's where the data is born. Everything downstream depends on this capture.
  • Tap to start — garage noise + UK tech voice play over each other, the way a real bay sounds.
  • At the populated checklist: "44 of 46 — two grey, because the AI is honest about what it missed. The SA catches those."
  • Warranty Admin point: "every OEM portal is a different UI. We push into all of them — the Admin stops typing the same MPI twice."
  • Transition to next slide: "so that's how the data gets captured. Here's what happens to it next."
Relay · Customer presentation
BOTTLENECK 02 · THE ADVISOR
Same pitch. Every customer. Every time.
HOW
  • Findings → priced quote: parts matched, labor codes mapped, both priced against your dealer's catalogs and markup rules — the invisible backbone
  • Quote packages into the customer pitch — photos, AI voiceover, plain-English explanation of each repair
  • Customer approves, declines, or asks to discuss — one screen
WHY IT WINS
  • No voicemail. No phone tag.
  • Advisor pitch variance → zero
  • HPRO up. Bay turn up. Red/Amber conversion up.
Speaker notes
  • Connective tissue: "same 44 findings you just watched get captured. Here's what the customer sees."
  • Tap "Play Your Vehicle Report" inside the iframe — let the audio carry it. Stay silent.
  • At the interactive Phase 4, pause. First line: "every customer gets this. Same pitch, same quality, every time."
  • Then: "how much of your current red/amber conversion gap is just advisor pitch variance?" — listen.
  • If he probes on "findings → priced quote": this is the hard part. DMS-integrated parts catalog lookup, current pricing, dealer-specific markup rules per category, labor codes against the vendor time guide. It's why our pitch is a quote, not an estimate. Every incumbent either punts this to the advisor or gets it embarrassingly wrong.
  • Do NOT go deep on on-device tech yet. That's slide 4.
Recite · Service Advisor portal
The middle layer

Advisor workflow — built to turn bays faster, not rubber-stamp.

What the SA sees at the counter. Not just a queue — a throughput accelerator.

Inbox · Westfield BMW
12 in queue · 2 sent back · 4 awaiting your review
9:42 AM
AB23 XYZ
BMW 3 Series 320d · Mineral Grey
Jake M. · 2 min ago · Recurring decline (Oct '25)
2R 2A 40G
YN54 GMJ
Ford Kuga 2.0 EcoBlue
Jake M. · 11 min ago
3R 19G
KL89 MNO
Vauxhall Corsa 1.2T
Mike T. · 24 min ago
5A 17G
FG67 HIJ
Toyota Yaris Hybrid
Jake M. · 38 min ago
22G
RV21 PQR
VW Golf GTD · 2022
Dave R. · 1 hr ago
1R 3A 18G
THE DIFFERENTIATOR
Tuned per workshop — not one-size-fits-all.
AI-assisted development means the SA workflow bends to each dealer's ops — queue rules, escalation policies, label language — not the other way around. Every incumbent is locked into what CDK or Xtime decided in 2014.
PLUS, THE TABLE STAKES DONE RIGHT
01
Voice-verified. The tech's walkaround audio plays inline — SA hears what was said before pushing to the customer.
02
Recurring declines flagged. If the customer declined this item 6 months ago, it surfaces automatically — Reengage firing in-workflow.
03
Rubber-stamping guard. Any advisor whose avg review time drops below 3 minutes gets flagged.
Speaker notes
  • 30–45 seconds total. Don't dwell.
  • Lead on the differentiator: "the middle layer isn't innovative by itself — this one thing is. AI-coded software means every dealer's SA workflow bends to their ops, not to whatever CDK or Xtime decided in 2014."
  • Run through the three table-stakes points fast — "voice-verified, recurring declines flagged, rubber-stamping guard."
  • Point #2 tees up Reengage. "That recurring-decline flag is Reengage firing in-workflow on the first visit."
Closing the Aftermarket Bleed
BOTTLENECK 03 · THE QUIETEST, BIGGEST MONEY

The declined job isn't dead. It's waiting for the right moment.

When a customer declines a $1,200 tire & brake job today, there's zero automated follow-up. That revenue walks to Discount Tire. Reengage doesn't just fire a dumb 30/60/90 cadence — it predicts the exact moment the customer is most likely to say yes.

1
Capture — automatically
On decline, Reengage saves the tech's photos, diagnosed prices, and the AI pitch Relay generated. No advisor action required.
2
Predict the right moment — per customer
An ML model decides when to fire, per item, per customer. Inputs: vehicle-physics decay curves, mileage trajectory, seasonal/weather signals, prior decline-to-buy latency, MOT proximity, recall windows.
3
Re-pitch, in context
Same delivery style as the in-drive Relay pitch — familiar to the customer. Content is adapted: wear estimates advanced for time passed, prices refreshed, urgency reframed for the trigger that fired ("your tread was 2.1mm in March — at your mileage, it'll be at 1.6mm before next week's cold snap").
4
Close the deal — zero-CAC
Customer taps to book. Work returns to your bay, not Discount Tire. Revenue you already diagnosed, recaptured without paying to acquire it.
"A platform of prediction models, not a campaign schedule. Each trigger is a tuneable model — and every closed-loop outcome trains the next one."
Signals the model weighs
Vehicle physics — tread wear curves, pad wear curves, fluid degradation
Mileage trajectory — projected odometer at any future date
Seasonal + weather — cold snaps, first frost, monsoon, tire-season transitions
Customer latency — prior decline-to-buy windows, payment-cycle hints
Regulatory calendar — MOT, emissions test, recall windows
Fallback — a naive 30/60/90 schedule only when signals are insufficient
Uses Recite's photos + Relay's pitch. Different trigger engine. Same revenue surface.
Speaker notes
  • Frame: "this is the quietest one, probably the biggest money. The version you'd pitch in a spreadsheet is 30/60/90 follow-up. That's not what we're building."
  • Money line: "every declined amber item is a prediction problem. The right moment is not 30 days, 60 days, or 90 days — it's the moment the customer is actually going to say yes. That's an ML problem and we're treating it like one."
  • Be honest: "Recite + Relay are shipping. Reengage is in design — the Zero-CAC recapture works on day one with the naive cadence; the prediction layer ships next."
  • If he asks about data: "every Relay impression + decision is a training signal. The product learns with every customer."
The platform layer — not yet built
THE INVISIBLE ENGINE

Reckon — turning findings into priced quotes.

The subsystem that makes Relay's pitch a quote, not an estimate. Relay and Reengage are only as good as the prices they quote — Reckon is where those prices come from.

WHAT IT DOES
  • Finding → Job Order translation
  • Parts lookup against OEM catalogs, DMS-integrated
  • Parts pricing with dealer-specific markup rules
  • Labor code mapping against vendor time guides
  • Labor pricing — dealer rate × time per op code
  • Bundling + tax + incentive rules (e.g. "while on the ramp")
  • Structured quote emission — consumed by Relay & Reengage
ARCHITECTURE
Recite
findings
→
Reckon
prices
→
Relay
presents
Reengage
re-pitches
STATUS
Not yet built. Reckon is the major build ahead — DMS adapters, catalog licensing, pricing-rules config. This is where help and investment are needed.
Speaker notes
  • Open honest: "we've shown you two shipping products and one in-design. Here's the fourth — a subsystem that isn't built yet."
  • Money framing: "the quote is the hardest thing in fixed ops software. Every DMS has tried. Most punt it to the advisor. We want to solve it as its own subsystem, so Relay AND Reengage both get accurate pricing on day one and years later."
  • Be explicit: "this is the biggest unknown, biggest risk, biggest opportunity. Building this right is where operator experience like yours matters more than capital."
  • Don't over-promise timelines. Say: "I can tell you what needs to happen. I need help deciding the DMS-integration order."
Why now
Why 2026, not 2025

Every enabler crossed the viability line in the last nine months.

CAPTURE
Recite
  • On-device STT fine-tuned for shop-floor noise and mixed accents
  • Compact LLM classifies Red/Amber/Green reliably — runs offline on a commodity phone
  • Full pipeline on-device — no Wi-Fi needed in the bay
PRESENTATION
Relay
  • Studio-tier TTS past the uncanny line — voice you can listen to without flinching
  • Narration cost < $0.20 per customer — and still falling
  • Mobile multimedia sync proven across consumer apps — template it into a product
RECAPTURE
Reengage
  • Per-customer retention prediction — bespoke ML last year, commodity now
  • Reuses Relay's TTS + delivery infra — zero new customer-facing surface to build
  • Cheap enough to fire per-customer, per-item — not one quarterly blast to the list
PRICING
Reckon
  • AI for judgment — labor codes, parts substitutions, quote composition
  • AI-built software for DMS plumbing — integrations in weeks, not quarters
  • Every dealer has a Polly from Parts — the clerk slammed all day fielding "what's the part number for…" Reckon takes the pressure off her.
Four products. One inflection window. None buildable a year ago.
Speaker notes
  • You've shown all four. Now the inflection-moment payoff: "these four products all became possible in the last nine months — because of the same AI wave."
  • Pace the four quadrants as: shipping now (top row), buildable next (bottom row).
  • The Polly line is the landing beat. Pause after reading it. Paul's parts department had a Polly. Say: "you know her. Every dealer has her. She's the bottleneck every tech curses. AI judgment plus AI-written DMS integrations take that pressure off her — and off your shop."
  • Close on the punchline: "none of these four were buildable a year ago. That window is the opportunity."
  • This sets up the close — transition: "so, with that said…"
Three products. One platform. Three bottlenecks closed.
Recite
Faster techs, perfect MPIs, OEM compliance — 10–15 min per RO.
SHIPPING
Relay
Same perfect pitch, every customer — HPRO up, bay turn up.
SHIPPING
Reengage
Automated BDC hunting lost revenue — Zero-CAC recapture, ML-timed.
DESIGN · 2026 H2
Reckon — the pricing-intelligence subsystem underneath. Not yet built. This is where help is needed.
Close the gap between what your techs find and what your dealership earns.
Is the juice worth the squeeze?
Speaker notes
  • Stop talking. Let the slide hang. He's done the math while you've been presenting.
  • If positive: "what's the smallest thing you'd want to see before getting more involved?"
  • If skeptical: "what part worries you most?" — either way, you learn.
  • Do NOT pitch fundraising, revenue, TAM, or pricing unless he asks. He's an operating partner — he'll open that door on his own schedule.