For service leaders & dealer principals
Three problems you deal with every week.
You already know what they cost you. This page is about whether Conduit actually moves them, on what timeline, and what it looks like in your shop.
The problems
Each one has a labor-rate number behind it.
Problem 01 · The repeat question
A tech asks the same question for the fourth time this month.
Your senior tech knows the answer. So does the manual. The new hire still walks across the bay to ask. Multiply by 30 bays, by 5 questions a day, by every dealer in your network.
Problem 02 · The ramp
The new hire takes 3 weeks to get useful — and 6 months to be trusted.
You hired them in March. They billed at 40% utilization through May. Your loaded labor rate doesn't care that they're learning. Neither does the customer waiting for the boat.
Problem 03 · The comeback
An expensive comeback. Wrong diagnosis, wrong part, angry customer.
Two hours diagnostic, parts ordered, vessel reassembled, customer picks up. Forty miles offshore the same fault returns. Now you eat the labor, the parts, the slip fee, and the relationship.
Before / After
Five scenarios from a real service week.
The left column is what your shop does today. The right column is what it does once Conduit is in the bay. No marketing hand-waves — specific scenarios, both columns.
Scenario
Tech needs a torque spec mid-job
Without
Walks to the office. Pulls the binder. Wrong year. Calls the senior tech who's elbow-deep in another engine. 12 minutes, two interruptions.
With
Asks Conduit on the iPad at the bay. Gets the spec, the page image, and the safety note in under 6 seconds. Cited to the manual.
Scenario
Multi-step diagnostic on an unfamiliar fault
Without
Senior tech runs it. The new hire watches when they can. The procedure lives in one head and walks out the door at retirement.
With
Diagnostic procedure surfaces step-by-step with the relevant TSB and warranty bulletin already attached. The new hire runs it. The senior tech reviews the result.
Scenario
Tech finds the documented procedure is wrong in practice
Without
They fix it the right way and tell nobody. Next tech hits the same wall. Knowledge stays in one head until that tech leaves.
With
Tech submits a correction in two taps. You review it in the queue. Once you approve, every tech in your shop gets the corrected answer the next time they ask.
Scenario
You want to know which procedures are slowing the bay down
Without
You guess. You ask the senior tech. They guess. You schedule training on the wrong topic.
With
The dashboard shows query volume by procedure, abstention rate by topic, and where techs got stuck. You schedule training on what's actually breaking.
Scenario
OEM ships a TSB that affects 200 of your units
Without
It lands in someone's inbox. Maybe gets distributed. Most techs find out on the third unit they touch.
With
The TSB is ingested. Every relevant query surfaces it inline with the answer. No bulletin chasing.
ROI — worked example
What 1 saved hour per tech per day looks like.
Numbers below model a 30-bay marine dealer. They are illustrative, not a benchmark. Drop in your own labor rate, tech count, and utilization assumption — the structure of the math doesn't change.
Assumptions — replace with yours
- Technicians in shop20
- Loaded labor rate (cost)$95 / hr
- Billable rate$165 / hr
- Hours saved per tech per day1.0 hr
- Working days per year240
Math
20 techs × 1 hr × 240 days = 4,800 recovered hours
4,800 hrs × $165 billable = $792K recovered revenue capacity
4,800 hrs × $95 loaded = $456K recovered labor cost
Directional outcome
~$790K capacity / yr
Per shop. At 30 bays. Before counting comeback reduction or ramp-time recovery.
What this is
A structure for the conversation. We'll rebuild this with your actual labor rate, tech count, and a measured time-saved number from your pilot.
What this isn't
A benchmark. We have one reference deployment in active use. We don't have a customer base big enough to publish averages. If anyone selling AI to your shop quotes a utilization gain to two decimals, ask them whose shop it came from.
What we'd measure in your pilot
Time-to-answer per query, query volume by tech, abstention rate, comeback rate before vs. after, ramp time of the next hire onboarded with Conduit in the bay.
First 14 days
What the rollout actually looks like.
Honest about what's automated, what's us-in-the-room, and where you have to give us a tech and a bay. No magic.
01
Day 1
Ingest your manuals
You send 2–3 PDFs (owner's manual, service manual, the most-referenced TSBs). We ingest, parse, chunk, embed. End of day you have a queryable corpus.
real
02
Day 2
Configure your tenant
Branding, equipment registry, role configuration (technician / service writer / leader), single sign-on stub. Your namespace is isolated from every other customer's data.
real
03
Day 3
Validation pass
We run a per-tenant eval set against your corpus — questions you'd actually expect a tech to ask. You get the pass-rate, the failures, and the documentation gaps before any tech sees it.
real
04
Days 4–7
Pilot in one bay
Pick one bay, one shift, one senior tech. They use Conduit on real jobs. We sit with them. Corrections go straight into the queue. End of week you have a feel for answer quality on real work.
real — needs your bay and your tech
05
Days 8–14
Roll out to the rest of the bays
Onboard the rest of the techs. Dashboard goes live for you. Correction queue, adoption view, knowledge-gap view. Weekly review with us.
real — onboarding is hands-on, not a self-serve flow yet
Two surfaces, one platform
What your techs see. What you see.
Same backend, two surfaces. Techs get answers. You get the management view of how the answers are landing.
What oil does the F300 require?
Specifications
Run engine to operating temperature before draining. Hot oil flows more completely. Use caution — surfaces will be hot.
What techs see
Asks a question in plain language at the bay
Gets a structured answer: spec table, procedure steps, safety warnings, cited pages
Taps a citation to see the original manual page
Submits a correction if the documented procedure is wrong in the field
What you see
Adoption: who's using it, how often, on what equipment
Knowledge gaps: where the system abstains, where techs ask the same thing twice
Correction queue: what your techs are submitting, ready for your review
Documentation health: which manuals are getting cited, which are silent
The leadership console
The views you'll live in.
Built for your weekly review. No charts for the sake of charts.
This week
847 queries · 31 techs active
Queries today
847
+12%
Active techs
31
+3
Citations served
2,104
+18%
Avg confidence
0.86
verified
Query volume · 14 days
+12% wk/wk
Adoption dashboard
Who's using it, on what equipment, how often. Spot the bay that hasn't logged in this week.
Knowledge gap detection
Where the system abstained. Where techs asked the same thing twice. The training topics that pick themselves.
TSB 2024-07 supersedes OM torque. Replace 38 N·m with 42 N·m on flywheel bolt.
F300 OM · p.184
Anode kit number changed for 2024 production: 6BG-11325-00 not 6CB-11325-01.
Parts catalog · 2024
Tell-tale water passage clears overheating in ~30% of cases — surface before thermostat replacement.
F300 SM · Cooling
ECU pairing requires key cycle off/on after firmware push — OM omits this step.
TSB 2024-07
Correction management
Your techs flag where the manual is wrong. You approve. Every tech in your shop gets the correction next time they ask.
Pilot
Pilot this on one bay.
Two manuals, one bay, one senior tech. Two weeks. At the end you have answer-quality data on your actual equipment and your actual team — and a real sense of whether the rest of the shop should follow.
1
2–3 manuals
Owner's manual, service manual, your most-referenced TSBs. PDFs are fine.
2
One bay, one senior tech
Someone who'll tell us when an answer is wrong. We need the signal.
3
A weekly half-hour
You see the dashboard, we triage corrections together, we tune.