01
SoH without a capacity test
Capacity estimated from normal daily cycles. No need to pull vehicles off the road for full discharge tests.
M-02State of Health AI
A health score that knows what is normal.
A battery at 78% can be fine or failing. It depends on its age, its chemistry and how it has been used. State of Health AI scores every battery against what is expected for a pack like it, and tells you which habits took the missing points.
SoH 71%Expected 84%Gap -13 ptsConf. ±2.4%
State of Health AI / EV-0447Confidence ±2.4% · 19 months data
What took the 13 points
- Fast charging to 100%-6.1%
- Heat exposure above 40 °C-4.3%
- Deep discharge below 10%-1.8%
- Cycle count for age-0.8%
Fleet SoH spreadEV-0447 marked
62%80%98%
Inside the module
What State of Health AI does
02
Expected health baseline
Each score is compared with a baseline for the same chemistry, age and duty cycle, so outliers stand out.
03
Cause breakdown
The gap between actual and expected health is split into causes: fast charging, heat, deep discharge, high C-rate.
04
Resistance tracking
Internal resistance trends alongside capacity, which often warn of trouble before capacity drops.
05
Fleet distribution
See the spread of health across the whole fleet and spot batches, suppliers or routes that age faster.
06
Confidence on every score
Scores show a confidence range based on how much clean data the battery has. Thin data means a wider band.
Current path
How it runs
- 1
Learn
Models are trained on charge and discharge curves across chemistries, then tuned to your packs during the first weeks.
- 2
Score
Every battery gets a State of Health score and an expected score, refreshed daily.
- 3
Explain
The difference is broken down by cause, so the next step is obvious: change charging, move routes or plan a replacement.
Rating plate
Specifications
Included on the plans shown in Pricing. Limits and integrations can be extended for large fleets.
Compare plans- Score refresh
- Daily, or after each full cycle
- Typical error
- ±2 to 3% SoH with cell-level data
- Baseline
- Chemistry, age, cycles, climate, duty cycle
- Outputs
- SoH, expected SoH, resistance trend, cause breakdown
Diagnostics
State of Health AI questions
Straight answers about data sources, hardware, accuracy and setup. Ask our team for anything else.
How is this different from the SoH my BMS reports?
Many BMS units report a fixed or slowly stepped SoH figure. We estimate it independently from voltage and current behaviour, and compare it against an expected baseline.
How long before scores are reliable?
Scores appear within days. Confidence tightens over the first three to six weeks as the model sees enough full and partial cycles.
Does it work for lead-acid batteries?
Yes, with a separate model. Lead-acid scores rely more on resistance and voltage sag than on coulomb counting.
Connected modules
M-03
Degradation Forecast
Predict remaining useful life and failure risk for every battery, months ahead, with confidence bands you can plan budgets around.
M-04Charging Insights
Find the charging habits that shorten battery life, by driver and depot, and roll out schedules that add months of service.
M-01Cell Monitor
Live readings from every battery management system and telematics unit in your fleet: pack and cell voltage, current, temperature and charge cycles in one view.
Charge complete · 0%