EOS · Fleet

WHOEVER DRIVES, EOS KEEPS THE RECORD.

Fleet records every vehicle, whether a person drives it, a system assists the person, or no driver is needed in a defined area. It derives what each vehicle costs a year and per kilometer, keeps the disagreements between sources side by side, shows which vehicles could be taken to an authority for driverless operation and which could not, and proposes with evidence. It proposes. It never drives, dispatches, commands or reroutes a vehicle, and it never changes a record it did not write.

Read this first. Whether a vehicle may operate without a driver in an area is the competent authority's decision, not EOS's. EOS records the permit an organization holds and shows a vehicle as operating without a driver only with a valid one on the record. A vehicle shown as a candidate is a question to take to the authority and the manufacturer, not a statement that it is able. Every saving is a scenario on a named assumption, and savings overlap, so EOS never adds them into one headline. Every figure in the film and in the platform walkthrough is on a synthetic fleet of fourteen vehicles.

Fleet in two and a half minutes

One fleet moving from people driving, to assisted, to driverless, what a kilometer costs each way, two odometers that disagree, permits that decide what is operating, a rising intervention rate routed to a named person, and one opportunity with its whole case.

What EOS holds and derives

Five kinds of record, entered by your people or delivered by a connector, and what EOS does with each.

Record	What it holds	What EOS derives
Vehicles	class, powertrain, automation level (SAE 0 to 5), acquisition, residual, insurance, distance, energy and maintenance per kilometer, driver cost, hours in service	what each vehicle costs a year and per kilometer, and how much of the hours it is available it is used
Events	services, faults, incidents, automated-driving interventions and minimal-risk stops, each with a date	counts and rates over the last year
Odometer sources	the reading telematics reports and the reading the service record holds	a disagreement when they differ beyond a tolerance. Both are kept; neither overwrites the other.
Service intervals	the odometer at the last service and the interval	which vehicles are past due, using the higher reading until a person settles which is right
Operating permits	the permit a vehicle holds for driverless operation: its identifier, area and expiry	a vehicle is shown as operating without a driver only with a valid permit on the record

What it proposes

Nine kinds of proposal, each from numbers you have recorded. A kind appears only where the records show it.

Proposal	Looks at	The figure is
Pool or retire under-used vehicles	hours in service against hours available	the fixed cost of the under-used vehicles
Service on schedule	unplanned maintenance and vehicles past due	the unplanned maintenance cost
Replace maintenance outliers	maintenance per kilometer against the fleet's own median	the cost above that median
Short-duty electric transition	daily distance of combustion vehicles within a planning range	energy and maintenance of those vehicles
Cut idling	the idle share and the energy cost	the energy spent idling
Charge off-peak		the energy cost of those vehicles
Driverless segment, to authority review	fixed-route share, utilization and driver cost	driver cost plus the fixed cost of fewer vehicles doing the same work. A question for the authority and the manufacturer, never a claim that a vehicle is able.
Restore or renew driverless permits	permits that are missing, expired or expiring	no dollar figure is claimed; it is a compliance and availability risk
Settle odometer disagreements	two sources that differ beyond a tolerance	no dollar figure is claimed; service timing depends on it

What every proposal carries

The data-backed insight and its evidence. The expected outcome as a low, expected and high figure beside the assumption that produced it. The impacts. At least two alternatives, each with its own insight, outcome and impacts, one of them doing nothing. The follow-on benefits, each labeled for what it is. A proposal for driverless operation also lists what it depends on: a permit, an approved system, a staffed remote-assistance operation and insurance terms.

Label	Means
OBSERVED	a recorded vehicle field or event, as someone entered it or a connector delivered it
DERIVED	computed from observed inputs by a rule the output names: annual cost, cost per kilometer, utilization, intervention rate
ASSUMED	how much of a cost base a lever could move. Named, bounded, shown beside the result and changeable per analysis
SCENARIO	a cost base times an assumption: a modeled future, never a promise
INFERRED	a follow-on benefit that is reasoned and not measured
RECOMMENDATION	the proposal itself
DECISION	what a named person decided, with the reason
OUTCOME	what a named person reports was realized

Where a person decides

A named owner, admin or controller accepts, declines or defers a proposal, with a reason, and later reports the realized saving with its evidence. A rising rate of interventions on a driverless vehicle is flagged against an internal review line and routed to a named person. The line is a screen for review, not a safety standard. An agent cannot decide, report or record vehicles.

How to use it today

Fleet is available in the platform application and through the EOS service. Every call is signed in, tenant-isolated and entered on your record. Sample records can be loaded into your own organization, each marked as a sample.

Where	What it does	Who
Fleet in the platform	Fleet, at /app#/fleet, shows the vehicles, cost per kilometer, autonomy readiness, the disagreements and every proposal with its alternatives, on a recorded synthetic run until you sign in.	Anyone in the walkthrough; your own records after sign-in
POST /fleet/vehicles	Records vehicles. Invalid ones are refused by name and the rest are accepted.	A named owner, admin or controller
POST /fleet/events	Records a service, fault, incident, intervention or minimal-risk stop.	A named person in the organization
POST /fleet/analyze	Derives costs and utilization, reconciles the sources, screens autonomy readiness and returns the proposals.	Any member of the organization
POST /fleet/decide	Accepts, declines or defers one opportunity, with a reason.	A named owner, admin or controller
POST /fleet/outcome	Reports the realized annual saving of an accepted opportunity, with its evidence.	A named owner, admin or controller
GET /fleet/ledger	Every proposal, decision and outcome, proposed set beside realized.	Any member of the organization

What it refuses, by name

Clause	When
fleet.person	An agent tries to record vehicles, decide or report an outcome. Only a named person does.
vehicle.invalid	A vehicle with a missing or impossible field. The fields are named and the other vehicles in the batch are still accepted.
assumption.range	An assumption outside zero to ninety percent, not a number, or not a known name.
vehicles.absent	There are no vehicles to analyze.
event.invalid	An event with an unknown kind, an unknown vehicle or no date.
decision.reason	A decision with no reason.
proposal.unknown	A decision about something that was not proposed to this organization.
outcome.undecided	An outcome reported for an opportunity nobody accepted.

What it does not do

It does not drive, dispatch, command or reroute any vehicle. It does not certify, approve or operate an automated driving system, grant or presume a permit, or determine safety. It does not change a record it did not write, place orders or approve spend. Two telematics readers, Samsara and Geotab, set odometer readings on vehicles already recorded; other vehicle data reaches it through the spreadsheet templates, a warehouse query or the registered operational-system push. It does not learn assumptions from outcomes by itself: the ledger sets proposed beside realized so a person can set the next assumption with evidence.