An AV-ready site scorecard evaluates a private property across ten weighted dimensions (access, capacity, hours, security, clearance, charging, service feasibility, risk, commercial terms, evidence) to determine if it can serve as a trusted ground-service node for autonomous fleets.
To evaluate whether a site is AV-ready, score it across ten dimensions: vehicle access, usable capacity, operating hours, security, vertical and turning clearance, charging potential, service feasibility, risk exposure, commercial terms, and evidence capability. Weight each dimension by how much it gates fleet operations, score it 1 to 5, and total it. A site is not “ready” because a vehicle can physically reach it. It is ready when a fleet can reserve it, enter it under known rules, complete a defined ground-service job there, and walk away with proof the session happened correctly.
What a fast site assessment looks at first
The fastest way to triage a property is to check the four dimensions that disqualify a site outright before scoring the rest. If any of these fails, the others do not matter.
| Gating dimension | Disqualifies the site if | Why it gates everything else |
|---|---|---|
| Access | No legal, repeatable way for an AV to enter and exit under a permission rule | A vehicle that cannot enter cannot stage, store, charge, or be recovered there |
| Clearance | Vertical height, turning radius, or grade blocks the target vehicle class | A Waymo Jaguar I-PACE, a Zoox shuttle, and a Nuro delivery pod have different envelopes |
| Hours | Site is locked during the fleet's overnight or recovery windows | Robotaxi ground work concentrates outside daytime peak; a 9-to-5 site is half-useless |
| Evidence | No way to prove a session happened (no records, no timestamps, no images) | Without evidence there is no SLA, no billing dispute trail, and no audit |
Everything below the gate is a matter of degree. The four gates are pass or fail.
What an AV-ready site scorecard is
An AV-ready site scorecard is a structured instrument for rating a private property against the operational requirements of an autonomous vehicle fleet. It replaces gut feel (“this lot looks fine”) with a weighted score across the dimensions that actually determine whether a fleet can use the space: who is allowed in, what they are allowed to do, whether the space is available, what rules apply, and whether the session can be proven.
It is not a parking suitability check and not a real-estate appraisal. A consumer parking lot is judged on price, proximity, and turnover. A ground-service node is judged on permission, capability, availability, and proof. Those are different questions with different failure modes.
Why scoring sites matters for autonomous fleets
Robotaxi operations are shifting from a depot-only model to a distributed one, and that shift makes site selection a primary operational risk. A depot is a known, controlled, owned environment. A distributed node is borrowed space with its own access rules, clearances, and liability. Picking the wrong node strands vehicles, breaks SLAs, and creates disputes no one can resolve because no one captured evidence.
Fleets are scaling into the hundreds and thousands of vehicles per metro. Waymo reported a fleet of roughly 3,000 vehicles serving more than 250,000 paid trips a week across its US markets as of early 2026, up from about 1,500 in spring 2025 (The Driverless Digest, “Waymo Stats 2025”; Waymo, “Scaling our fleet,” 2025). At that scale, the driver disappears but the ground work does not: vehicles still need to stage, store, charge-adjacent queue, get cleaned, get inspected, and get recovered when something goes wrong. Every one of those jobs has to happen somewhere physical, and “somewhere” is a scored decision, not an accident.
The cost of a bad node is concrete. A vehicle that cannot exit a site after hours is out of service until a human arrives. Remote assistance can reach a stuck robotaxi within minutes, but field recovery requiring a human to physically attend the vehicle can stretch much longer: documented incidents in San Francisco and Redwood City put on-scene recovery at roughly 30 to 50 minutes per event (TechCrunch, Mar 2026; KQED). Multiply that across a fleet and a few badly chosen sites become a daily reliability tax.
How the scorecard works
The scorecard rates ten dimensions on a 1-to-5 scale, multiplies each by a weight, and sums to a weighted total. Weights reflect how hard a dimension is to fix and how completely it gates operations. Access and evidence are weighted heavily because a failure there is fatal. Commercial terms are weighted lower because they are negotiable.
The 1-to-5 scale uses explicit anchors so two assessors score the same site the same way:
- 1: Disqualifying. The dimension blocks fleet use entirely.
- 2: Major gap. Usable only with significant remediation by the property.
- 3: Workable. Meets the minimum with documented constraints.
- 4: Strong. Clean fit with minor caveats.
- 5: Ideal. Purpose-fit, low-friction, low-risk.
A site that scores 1 on any gating dimension (access, clearance, hours, evidence) is failed regardless of its weighted total. The weighted total only ranks sites that already clear the gates.
Who needs an AV-ready site assessment
Four roles use this scorecard, for different reasons:
- Property and real-estate owners use it to find out whether the asset they already own can earn revenue as a ground-service node, and what it would take to qualify.
- Parking operators use it to convert daytime-empty or overnight-empty capacity into off-peak fleet support, scored against real AV constraints instead of consumer-parking instincts.
- Fleet-ops partners (the teams running ground operations for an autonomy company) use it to qualify candidate sites before signing access agreements.
- AV infrastructure teams use it to standardize how a network of distributed nodes is evaluated, so site quality is comparable across a metro.
Example: scoring a mid-block parking structure
A fleet-ops partner is evaluating a 180-space downtown parking structure for overnight staging and charging-adjacent queueing. The illustration below shows how the scorecard resolves a real trade-off: good access and hours, but a clearance problem.
| Dimension | Raw score (1-5) | Note from the assessment |
|---|---|---|
| Access | 5 | Gated entry, single controlled ingress, geofence-able approach |
| Capacity | 4 | 40 contiguous spaces allocable overnight |
| Hours | 5 | 24/7 access with after-hours code |
| Security | 4 | Staffed lobby, full camera coverage, lit |
| Clearance | 2 | 6'8" deck clearance blocks taller sensor-stack vehicles |
| Charging | 3 | Power capacity present, no chargers installed yet |
| Service feasibility | 3 | Room for cleaning and inspection, no wash bay |
| Risk | 4 | Low flood/crime exposure, clear liability owner |
| Commercial terms | 4 | Operator open to revenue share |
| Evidence | 4 | Cameras and access logs exportable for session proof |
The clearance score of 2 is the headline. The structure is excellent for a sedan-class fleet and unusable for a taller shuttle or roof-stacked sensor platform. The scorecard does not say “good” or “bad.” It says good for this vehicle class, scored, with the constraint named in writing. Code minimum clear height for parking structures is 7′0″ (2,134 mm), but real-world garages range from about 6′8″ in older downtown decks to 8′0″ in modern facilities (UpCodes, “Clear Height”; THA Consulting, parking design standards). A sedan-class AV such as a Jaguar I-PACE clears most decks, but a taller shuttle or roof-mounted sensor stack does not, which is exactly the trade-off this dimension scores.
What XoomPark does and does not do with the scorecard
XoomPark is the AV ground-services layer. It uses the scorecard to qualify sites and then coordinates everything that happens after qualification: reservation, access, workflow, SLA, evidence, and audit. The scorecard is the front door; the coordination is the building.
| XoomPark does | XoomPark does not do |
|---|---|
| Score and qualify private sites as ground-service nodes | Own or operate chargers or an EV charging network |
| Record private-site access rules and permissions | Perform vehicle maintenance, washing, or repair itself |
| Hold reservation and session records per visit | Dispatch the fleet or replace the fleet operator |
| Run check-in / check-out workflow and exception handling | Map the site or tell the AV how to drive |
| Track SLAs and capture evidence for audit and billing | Certify AV safety or act as an autonomy company |
The distinction that runs through all of it: the property defines permission, the fleet validates capability, and XoomPark coordinates reservation, access, workflow, SLA, evidence, and audit. Mapping tells a vehicle where it may be able to drive. The scorecard and the layer above it tell a fleet where it is allowed to go, what it may do there, whether the space is free, and what proves the session was clean.
Original analysis: the ten weighted dimensions, published in full
Most AV-readiness content stops at “make sure your site is accessible and secure.” We publish the actual instrument. The table below is the scorecard: ten dimensions, the weight we assign each, what a high score looks like, and the gating dimensions marked. The weights are XoomPark's working model for qualification scoring, presented as an illustrative methodology rather than a fixed industry standard or a figure measured from operating data. They reflect how much each dimension gates fleet use and how hard it is to remediate.
| # | Dimension | Weight | Gating? | What a 5 looks like |
|---|---|---|---|---|
| 1 | Access | 18% | Yes | Legal, repeatable, geofence-able entry/exit under an explicit permission rule |
| 2 | Capacity | 12% | No | Enough contiguous, allocable space for the target job during the target window |
| 3 | Hours | 12% | Yes | Available in the fleet's overnight and recovery windows, not just business hours |
| 4 | Security | 10% | No | Controlled access, full camera coverage, lighting, staffing or monitoring |
| 5 | Clearance | 12% | Yes | Vertical height, turning radius, and grade clear the target vehicle class |
| 6 | Charging potential | 8% | No | Installed chargers or sufficient power capacity to add them charging-adjacent |
| 7 | Service feasibility | 8% | No | Physical room for the ground-service job (clean, inspect, stage, hold) |
| 8 | Risk | 8% | No | Low flood/crime/liability exposure, single clear responsible party |
| 9 | Commercial terms | 6% | No | Owner willing to contract access on workable revenue or fee terms |
| 10 | Evidence | 6% | Yes | Exportable logs, timestamps, and imagery sufficient to prove a session |
Methodology note. Total weighted score runs 1.00 to 5.00. A working qualification threshold in this model is a weighted total of 3.5 or higher AND no gating dimension scoring below 3. Two sites with the same total are not equal if one carries a gating constraint: the gate is checked first, the total ranks second. This two-stage logic (gate, then rank) is the part a single averaged star rating gets wrong, because an averaged score lets a strong site hide a fatal access or clearance failure inside a high mean.
When you do not need a site scorecard
If you run a single fixed depot and never operate off-site, you do not need a distributed node-scoring instrument yet. Your one site is already known, controlled, and owned: you assess it once and you are done. The scorecard earns its keep when you have many candidate sites, borrowed space with varying rules, or a metro you are trying to cover with distributed capacity. It is also the wrong tool if you are choosing a consumer parking lot for human drivers: that decision is about price and proximity, not permission and proof. And it does not replace a structural, electrical, or legal inspection. It tells you which sites are worth inspecting.
Frequently asked questions
How do you evaluate whether a site is AV-ready?
Score it across ten weighted dimensions: access, capacity, hours, security, clearance, charging potential, service feasibility, risk, commercial terms, and evidence. Gate on access, clearance, hours, and evidence first (any failure there disqualifies the site), then rank the rest by weighted total. Readiness means a fleet can reserve the site, enter under known rules, complete a defined job, and prove the session.
Is an AV-ready site scorecard the same as a parking suitability check?
No. A parking check rates price, proximity, and turnover for human drivers. A site scorecard rates permission, capability, availability, and evidence for an autonomous fleet. A lot can be a great parking lot and a failing ground-service node if it locks overnight or cannot prove a session happened.
What disqualifies a site immediately?
Four gating failures: no legal repeatable AV access, clearance that blocks the target vehicle class, hours that exclude the fleet's overnight and recovery windows, or no way to capture evidence of a session. Any one of these fails the site regardless of how well it scores elsewhere.
Does the scorecard certify that a site is safe for autonomous vehicles?
No. The scorecard qualifies a site for ground-service use and records its constraints. It does not certify AV safety, inspect structures, or validate that the vehicle can drive the space. The fleet validates driving capability; the property defines permission; XoomPark coordinates the records around both.
Related pages
AV ground services: the full ground-operations layer →
What an AV-ready ground node is →
AV-ready SLA evidence and session proof →
How parking operators turn empty capacity into fleet support →
How real-estate owners qualify property as a ground node →
Private-site AV access rules and permissions →
The AV-ready staging and evidence pilot →
Request an AV-ready site scorecard
Have a property, lot, or structure you think could serve an autonomous fleet? Request an AV-ready site scorecard and get your site rated across all ten dimensions, with the gating constraints named in writing.