06Roadmap
6 of 7 · 10 minThe build sequence, and the gates you do not skip.
Eighteen months from audit to platform, with three gates, the honest cost, and a 90-day checklist you can tick off between meetings.
- What this is
- The eighteen-month build sequence, the three gates, and the first ninety days.
- Why it matters to you
- Most stores fail here by buying layer four while sitting on a broken layer one.
In this module you will
- ✓Sequence the 18-month build without skipping a layerWorked when · Your plan starts with audit and ownership, not a deployment date.
- ✓Apply the three gates before buying, deploying and scalingWorked when · Each gate has a person who can say no.
- ✓Commit to the first 90 days in writingWorked when · The 90-day checklist has dates and owners against every line.
A scenario from 2028
That was not an outlier. That was a preview.
A customer named Marcus Chen walked in to buy an Expedition. Before the F&I manager opened the menu, Marcus said: “I have already been pre-approved at 5.1%. I ran your GAP product against three alternatives and the expected value does not work for this vehicle’s loss history.” He had a printout, generated that morning by an AI agent.
The problem did not disappear overnight. It graduated. And the dealer did not graduate with it in time. The window to act: 18 months.
Vehicles / mo
1,104 → 487
Front-end gross
$1,840 → $214
F&I gross / unit
$2,490 → $1,380
Oct 2025 → Oct 2028
The build sequence
Three phases. Eighteen months.
Foundation
- →Audit your data, and know your duplication rate
- →Identify your tech-first employee and promote them
- →Pick one workflow (service scheduling) and start there
First Wins
- →Service scheduling agent in production
- →Declined service follow-up automation (4-8x ROI in year one)
- →F&I pre-qualification briefings running before appointments
Intelligence Layer
- →The AI nucleus stops being tools and becomes a platform
- →Which customers are 90 days from lease end?
- →Which advisors convert declined service at twice the peer rate?
The three gates
You do not advance on the calendar. You advance when the gate opens.
Most failed builds moved to the next phase on schedule with the last one unfinished. If a gate will not open, do not push past it. Fix what is holding it shut.
Gate One
- →Duplication rate measured, and under 10 percent
- →Full data export in your hands, not promised
- →One written definition of customer, signed by leadership
Gate Two
- →The workflow documented end to end, every step, every owner
- →A named operator with four protected hours a week
- →Winning defined in writing, and the number that proves it
Gate Three
- →First agent has run 90 days with a measured result
- →One vendor line item canceled and the savings booked
- →The next three workflows documented and waiting
The honest number
Nobody fails on the budget. They fail on the hours.
$115K
Cash, year one.
6 hrs
Controller, weekly, for 90 days.
4 hrs
Operator, weekly, protected.
12
Workflow sessions, year one.
Every vendor quotes you the first number. None of them quote the other three, and the other three are why implementations stall. What you already pay vendors is $150K to $400K a year, and none of that spend buys anything you keep.
This fails in predictable ways
Every one of these has already been avoided by somebody.
01
Bought the tool before running the data audit. The tool worked fine. The data underneath it did not.
02
No named owner, so the documentation died inside thirty days.
03
Deployed an agent onto a workflow nobody had written down. It ran the bad process faster.
04
Found the tech-first employee and never protected the time. New title, same job, nothing shipped.
05
Leadership mandated AI without touching the systems underneath. The mandate made activity, not results.
06
Ran a pilot with no written definition of winning. It never failed. It also never ended.
07
Treated month three as a milestone instead of a gate, and moved on to agents with dirty data anyway.
The receipts
Real stores. The last twelve months.
$1.3M
Email spend, gone.
A 45-store group built email on its own data and canceled the marketing vendors. Same volume, full attribution, zero middlemen.
3 days → 15 min
Month-end close.
A six-store Dallas group automated close-out on owned data. The office spot-checks the output instead of grinding through it.
1,200 calls
Before the storm hit.
A Midwest group pulled every declined tire and brake job ahead of the first snowstorm and called each customer. The service line spiked off the chart.
$0
Equity mining. Dashboards. Data fees.
Once the data is yours, the tools reselling it get canceled: equity mining in month one, analytics dashboards next, then every per-vendor fee.
None of them added headcount. They added infrastructure.
Your 90-day plan
Twelve moves. Tick them off between meetings.
01 Data
02 Workflows
03 AI
04 Agents
Progress is saved in this browser only.
Seven questions
Answer these honestly.
They will tell you which version of the dealership you are actually building.
Q1
What percentage of your gross profit depends on information the customer no longer needs you for?
Q2
How many of your service lane repair orders could be resolved by an OTA update by 2028?
Q3
Does your customer's primary digital relationship go through you, or through your OEM?
Q4
Who in your building is already building things without being asked?
Q5
If information asymmetry in your market were eliminated tomorrow, what is left that you would still get paid for?
Q6
Is the AI you are buying product evolution or true innovation?
Q7
Which version of the dealership are you actually building, based on the decisions you made in the last 90 days?
Before you leave
One commitment. Not sixteen.
- 01
Write down the one thing
Not the list. One line. The single move you make before this group meets again. Sixteen assignments produce zero. One produces one.
- 02
Put a name and a date on it
Yours or someone in your building, and a date before the next meeting. A commitment with no name on it is a wish.
- 03
Report the number at the next meeting
The number, not the effort. Duplication rate, workflow count, operator's name. The group holds you to it.
Key takeaways
- →Gates exist so you stop buying ahead of your foundation.
- →Months 1-3 are audit and ownership, not deployment.
- →The 90-day checklist is the deliverable for the next meeting.
0 of 7 complete