One platform. Every marque in the group.
AiR is built once and dressed per brand. The hardware, the four-channel logic and the pressure derivation engine are identical across Sytner's premium and luxury estate. Only the brand layer, the placard data and the owner language change.
Group brands
0
Marques represented across the Sytner premium and luxury estate
Portals live
0
Marque portals built and demonstrable across the group
Retail sites in scope
0
Indicative UK premium and luxury retail sites across the group
Economy lost per PSI
0.0%
Fuel or range penalty for every 1 PSI under target
Discussion draft. Sytner Group, its marques and all co-branded visuals are proposal concepts. No endorsement is implied and commercial figures are indicative and non-binding.
01 Live portals
Built, branded and demonstrable today
Every marque in the group now has a complete portal: brand-scoped dashboard, owner app, pilot plan and commercial model, each in its own identity.
03 Fleet at scale
What backing this looks like across the estate
Move the estate size to see the capital envelope, the verified checks generated and the aftersales value attributable to the bay. Indicative planning figures, not a quotation.
Sites deployed
24
Capital envelope
GBP 567,600
Units, brand livery and installation across the deployed estate
Verified checks a year
51,840
At 180 modelled sessions per site per month
Advisories raised
10,886
At a 21% conversion from bay session to identified work
Aftersales value
GBP 2,024,796
Advisories multiplied by an average GBP 186 job value
Managed alternative
GBP 158,112 / yr
If the group takes the machines as a managed service instead of capital
Indicative payback
3 months
Capital divided by attributable annual aftersales value
04 Return, social value and brand impact
What the group gets back, in money and in reputation
Financial return, road-safety and environmental value, and the brand outcomes that follow. Move the estate size to scale every figure. Indicative planning numbers, not a quotation.
Estate modelled
24 sites
Typical ROI, five years
1684%
Five years of attributable aftersales value against the capital envelope
Indicative payback
3 months
Capital divided by attributable annual aftersales value
Annual value per site
GBP 84,367
Advisory value attributable to a single bay each year
Net first year, managed
GBP 1,866,684
Attributable value less the managed service fee, no capital outlay
Breakdowns avoided a year
332
Modelled from tyre-related breakdown rates applied to verified checks
CO2 avoided a year
213 t
At 0.2% economy lost per PSI under target across the modelled correction
Fuel and energy saved
88,128 l equiv.
Owner-side saving from correctly inflated tyres between services
Tyres diverted from landfill
166
From a modelled 12% extension in tyre life
Tyre particulates avoided
933 kg
Non-exhaust emissions reduced by removing shoulder scrub from under-inflation
Owners given a safety record
12,960
At a 90 day Verified Care cadence, one record per owner per quarter
01 Brand awareness
+38% earned reach
A branded four-channel bay on the forecourt is a physical marketing asset. Modelled uplift combines forecourt impressions, owner app opens and the press value of being first in the group to publish a verified tyre-safety number.
02 Customer retention
+9 pts service retention
A verified check every 90 days creates four owner touchpoints a year between services. Retailers keep the relationship, and the warranty record only stays claimable while the checks continue at the retailer.
03 Customer satisfaction
+11 NPS points
Owners leave with four measured pressures, four targets and the reason for each correction rather than a generic placard number. Transparency at the bay is the single largest driver in the modelled satisfaction uplift.
04 Trust and advocacy
72% would recommend the bay
Precision borrowed from professional motorsport, applied to a family car, is a story owners repeat. Advocacy is measured on the referral of the bay itself, not the last vehicle purchase.
05 Employer and social licence
Board-ready safety number
One auditable compliance figure across every marque supports ESG reporting, local road-safety partnerships and the recruitment story that the group invests in preventative safety rather than reactive repair.
05 One record, five audiences
How the group-wide tyre record is read
Every verified check writes once and is read at five levels, from the technician holding the hose to the board publishing a safety number.
Group board
Level 01One compliance number across every marque: the share of vehicles that arrived outside tolerance, by brand, by region and by quarter. This is the number that makes a road-safety claim defensible in public.
Brand director
Level 02Marque-level pressure drift, seasonal behaviour and premature-wear signals, mapped against the homologated fitments that manufacturer sanctions. Evidence for the manufacturer conversation, not marketing copy.
Aftersales director
Level 03Bay throughput, advisories raised, conversion into booked work and the revenue attributable to the bay that raised it. Measured per site rather than claimed per network.
Retailer principal
Level 04Daily session volume, queue length, consumable state and the vehicles that left with a flag needing follow-up. One screen, one bay, one day.
Technician
Level 05The verified record for the car in front of them: four measured pressures, four targets, the load declared and the correction applied, timestamped and signed.
06 Edge compute, AI and impact
What onboard AI adds for Sytner
Every deployed unit carries its own accelerator. That turns the estate into a distributed AI network that improves with use, keeps working when the line drops, never pools customer data, and produces a safety and carbon record Sytner can publish.
retail bays deployed
24
Pilot scale is 24 retail bays. Full modelled scope is 140. Compute, social value and carbon all scale from this one input.
Aggregate peak inference
6,600 TOPS
Jetson AGX Orin at tier F2 - High-value site / backbone anchor, INT8 peak per published module specification
Reserved for safety work
2,310 TOPS
Held back for local tyre inference and the guided check before any capacity is offered out
Shareable capacity
4,290 TOPS
Sovereign UK inference capacity offered on a host-share split once safety headroom is reserved
Federated rounds a year
25
Model updates aggregated across the cohort. Raw session data never leaves the node
Host-share revenue
GBP 61,632 / yr
Modelled host share at the assumed duty cycle. Indicative planning figure, not a quotation
Estate energy draw
12,624 kWh / yr
Compute envelope only, at the published annual figure per module tier
01 Every bay becomes a compute node
The unit in the bay is not a pump with a screen. It carries a Jetson class accelerator running tyre condition inference, wear classification and pressure derivation locally, in the seconds the owner is standing there. No round trip to a cloud region, no dependency on a retailer's broadband.
02 The model improves without pooling customer data
Raw session data stays on site. Only aggregated model updates leave, so the network learns from every check across the estate while each retailer's customer records never move. Federated learning gives the group a tyre model no single dealership could train alone.
03 Uptime that survives a dead line
If cellular and site broadband both fail, the bay keeps working. Inference is local, the record queues locally, and the sovereign mesh carries the safety-critical summary out over LoRa and Reticulum until the primary link returns.
04 A sovereign UK data position for procurement
Personal data is processed at the edge and stays in the UK. That answers the procurement, DPIA and manufacturer data questions before they are asked, which is usually the slowest part of approving new forecourt technology.
05 Second income from hardware already installed
Idle accelerator capacity, above a reserved safety headroom, is offered as sovereign UK inference capacity. The estate earns from compute it already owns, on a host-share split, without adding a single square metre of data centre.
06 AI that produces a record, not a guess
Every inference is bound to four measured pressures, four derived targets and a timestamp. That makes the AI auditable. It is the difference between a model the board can publish against and a model nobody can defend.
Verified checks a year
51,840
At 180 modelled sessions per retail bay per month
Breakdowns avoided a year
332
Modelled from tyre-related breakdown rates applied to verified checks
Owners given a safety record
12,960
At a 90 day Verified Care cadence, one stamped record per owner per quarter
Tyres diverted from waste
166
From a modelled 12 percent extension in tyre life once pressure is held on target
Tyre particulates avoided
933 kg
Non-exhaust emissions reduced by removing shoulder scrub caused by under-inflation
Fuel cost avoided
GBP 122,366 / yr
Money back to households and fleets at the blended UK pump price, not revenue to the operator
The same inference results, expressed as the outcomes a board, a council or a journalist asks about: tyre health held on target, roadside incidents that never happen, and safety awareness created at the point of measurement.
Tyre health uplift index
+17 pts
17% of modelled full estate
Modelled index movement per vehicle held on target across a year of checks, on a 0 to 100 condition scale
Roadside incidents avoided
332
17% of modelled full estate
At 6.4 tyre-related incidents avoided per 1000 verified checks
KSI exposure reduced
2.2 events
17% of modelled full estate
Killed or seriously injured exposure, modelled from the published tyre-related casualty share. Modelled, never a measured casualty saving
Roadside hours returned
531
17% of modelled full estate
Hours not spent stationary on a live carriageway or in a layby waiting for recovery
Recovery miles avoided
8,958
17% of modelled full estate
Fewer attended call-outs means fewer miles driven to attend them, a second carbon line
Community safety conversations
21,254
17% of modelled full estate
At 41 explained results per 100 checks. Awareness created at the point of measurement, not by campaign
Owners holding a verified record
12,960
17% of modelled full estate
One stamped record per owner per quarter at the Verified Care cadence
Tyres diverted from waste
166
17% of modelled full estate
From the modelled 12 percent life extension once pressure is held on target
Every figure above starts as an inference run on the node: four measured pressures, a derived target, a wear classification. The outcome is the consequence of the compute, not a separate marketing model.
The uplift is concentrated on the vehicles least likely to be checked anywhere else. That is what makes the number a public good rather than a marketing metric, and what makes it publishable with a local authority.
No casualty saving is claimed as fact. KSI exposure is a modelled reduction in risk, derived from published casualty shares, and it is reported that way in every document.
Outcome figures are modelled from published per-check factors applied to the modelled session volume at the selected scale. They describe expected outcomes, never a measured casualty, breakdown or emissions result.
Estimated CO2e avoided at this scale, across a conservative to optimistic band, with every assumption on screen and the numbers downloadable for a sustainability team to check line by line.
Session volume comes from the scale selector above. Everything else is yours to challenge.
Diesel share of fleet
38%
Checks that find a correction
57%
Litres corrected per check
1.70
Conservative case
44 tCO2e
after the uncertainty haircut
Estimated before haircut: 68 tCO2e
Fuel demand avoided: 30,140 litres
Social value of carbon: GBP 12,086
Fuel cost avoided: GBP 41,849
Central case
74 tCO2e
after the uncertainty haircut
Estimated before haircut: 114 tCO2e
Fuel demand avoided: 50,233 litres
Social value of carbon: GBP 20,143
Fuel cost avoided: GBP 69,748
Optimistic case
107 tCO2e
after the uncertainty haircut
Estimated before haircut: 165 tCO2e
Fuel demand avoided: 72,838 litres
Social value of carbon: GBP 29,208
Fuel cost avoided: GBP 101,135
These are avoided emissions estimates, not carbon credits. AiR issues a Measured Intervention Record and an Avoided Emissions Estimate today, and an evidence certificate on top of it. The words credit, offset, carbon neutral and net zero are not used for AiR units before a recognised registry has issued them.
Fuel demand avoided
88,128 litres
Litres not burned, modelled from the correction applied per verified check
Avoided emissions, estimated
204 tCO2e
At the 2025 UK conversion factors, 2.069 petrol and 2.571 diesel kgCO2e per litre, blended
After uncertainty haircut
133 tCO2e
A 35 percent conservative haircut applied before anything is put forward as inventory
Social value of that carbon
GBP 36,281
At the GBP 273 per tCO2e central appraisal value. A welfare value, never shown as revenue
Hierarchy level reached
Level C
Measured Intervention Record, Avoided Emissions Estimate and AiR Carbon Impact Certificate. Level D, a registry-issued credit, requires validation and verification first
Tradeable today
No
Nothing above is a credit or an offset until a recognised registry issues it under an approved methodology
These are avoided emissions estimates, not carbon credits. AiR issues a Measured Intervention Record and an Avoided Emissions Estimate today, and an evidence certificate on top of it. The words credit, offset, carbon neutral and net zero are not used for AiR units before a recognised registry has issued them.
01 A publishable safety number
Group compliance, drift and correction rates, generated automatically rather than surveyed. This is the number that makes a road-safety statement defensible in public and quotable by the press.
02 First mover on responsible roadside AI
Edge inference, data kept local, no customer data pooled. A group that can say that is positioned ahead of every competitor still shipping data to a foreign cloud region.
03 ESG reporting with an audit trail
Avoided litres, avoided emissions and avoided breakdowns tie to individual verified records, so the environmental claim is traceable to a session rather than to a spreadsheet.
04 Local partnership material
Regional road-safety partnerships, council and police campaign tie-ins and school engagement all become possible once the estate holds real local data on tyre condition.
Node counts, TOPS totals and host-share revenue are planning assumptions against published module specifications. Social and carbon figures are modelled from per-check factors, not measured outcomes, and are labelled as such wherever they appear.
07 The group case
Why this is a group decision, not a dealership one
One platform, many marques
01The hardware, the four-channel logic and the pressure derivation engine are identical across every brand. Only the brand layer, the placard data set and the owner language change, so a second marque costs a fraction of the first.
Group-level tyre data
02Every verified check writes into one group record. Sytner sees pressure compliance, seasonal drift and premature wear signals by brand, by site and by parc rather than one dealership at a time.
Aftersales pull-through
03A pre-journey bay gives owners a reason to visit between services. Advisories raised at the bay route to the retailer that raised them, which is measurable service conversion rather than a marketing claim.
Manufacturer alignment
04Each marque keeps control of its own brand rules and pressure data. AiR never publishes a target the manufacturer has not sanctioned, which makes brand approval a review rather than a negotiation.
Staged rollout
05Ferrari and Aston Martin first, as the marques where precision is already part of the ownership language. Premium volume brands follow once throughput and reliability are proven on low-volume, high-attention sites.
Discussion draft. Sytner Group, its marques and all co-branded visuals are proposal concepts. No endorsement is implied and commercial figures are indicative and non-binding.