Intelligence Applied

What changes when the battery is no longer unknown.

One decision per battery per week. For every stakeholder in the Indian EV chain. Automatically.

4 stakeholders. One platform.
Weekly. One decision per battery. Automatically.
For the first time —
Battery OEM
Sees how their design performs in Indian field conditions.
↳ Not the lab.
NBFC / Lender
Knows the health of the asset they lent against. Before default.
↳ Not the invoice. Not the court.
Fleet Operator
Knows which vehicle stops — six weeks before it stops.
↳ Not after.
Regulator / Vehicle OEM
Sees the gap between specification and Indian field reality.
↳ Not the lab report.
What changes

Four stakeholders. One intelligence chain.

Battery OEM
Have
Lab test results. Warranty claims at month 14. No field visibility.
What changes
Field degradation patterns by pack × city × use case. Attribution in 90 seconds. Design feedback from real Indian stress conditions.
Fleet Operator
Have
Range today. Driver complaints. Reactive service calls.
What changes
Which battery needs action this week — and why. Replace vs charger swap. Six weeks of advance warning. 73% cost saving on false replacements.
NBFC / Lender
Have
Loan amount disbursed. Monthly payment status. Asset invisible between disbursement and default.
What changes
Battery health grade A/B/C/D updated weekly. Breach probability 6–8 weeks early. Collateral mark-to-market. Default from the asset, not from credit.
Regulator
Have
Lab-derived range specifications. BMS alerts calibrated for non-Indian conditions.
What changes
Field-validated range data. Battery passport written weekly. Regulatory readiness before the mandate arrives.
enerlystAI · Agentic processes

The agent works. You approve.

Three processes running in production. Physics-grounded inputs. One decision per asset per week. Human-in-loop approval on every action.

NBFC / Lender
NPA Management
Battery health deteriorates weeks before the EMI fails. The agent detects it. The NBFC acts before the default — not after.
enerlystAI · PROCESS 01 · WEEK 47
ASSET ···············
BAT · LFP · 278
GRADE ···············
D — CRITICAL
BREACH PROBABILITY ··
67% within 8 weeks
ATTRIBUTION ·········
38% operator · 29% OEM
21% use case · 12% calendar
ACTION ··············
Initiate collections contact.
Do not wait for EMI miss.
< 4 min · zero manual intervention
LIVE
What happens next
01Agent generates credit committee report — battery health trajectory, attribution split, breach probability curve. Format: PDF-ready structured output.
02If no contact by Week 49 — agent escalates to collections priority queue. Breach probability recalculated with updated telemetry.
03If telemetry shows battery health recovering — agent downgrades alert. Grade D → C. Collections contact deferred. Human notified.
Human approves every escalation. Agent prepares. Human decides.
Battery OEM
Warranty Attribution
Claim arrives. OEM disputes. Months of negotiation. The agent resolves it in 90 seconds with physics-grounded evidence.
enerlystAI · PROCESS 04 · CS-04
ASSET ···············
Pack3001 · Month 14 failure
ATTRIBUTION ·········
34% operator charging
28% OEM design
22% IoT / commissioning
16% calendar aging
VERDICT ·············
PARTIAL
OEM LIABILITY ·······
Reduced 61% vs full claim
SHAP R² ·············
0.816 · auditable
90 seconds · evidence pack ready
LIVE
The evidence pack
01SHAP decomposition rendered as a signed PDF — each factor, its contribution, the corroborating signal chain. Legally auditable.
02Commissioning data cross-referenced: cell spread at Week 1 vs fleet P75. Deviation flagged as OEM design factor — not operator abuse.
03Settlement recommendation: OEM liability reduced 61% vs full claim. Operator contribution acknowledged. Dispute resolved before it begins.
Attribution is not an opinion. It is a physics-grounded calculation with a confidence bound.
Fleet Operator
Predictive Service
Battery degrading. No BMS alert. No driver complaint. The agent detects the regime change and prescribes the cheapest intervention first.
enerlystAI · PROCESS 04 · WEEK 47
ASSET ···············
BAT · LFP · B
FIELD RANGE ·········
68km · declining 1.2km/week
PELT SIGNAL ·········
Regime change · Week 43
FORECAST ············
Floor breach · 6 weeks · 82.4%
ACTION ··············
Charger swap first.
Saves 73% vs replacement.
Review Week 49.
Weekly · PELT + CUSUM + Conformal
LIVE
The intervention sequence
01Week 47: Charger swap recommended. Fleet manager notified. Cost: ~₹3,000. Saves 73% vs full battery replacement.
02Week 49: Agent checks telemetry. If cell spread recovering — close the alert. Vehicle back to STABLE. Replacement avoided.
03Week 49: If no improvement — replacement confirmed. Scheduled during vehicle downtime. Driver not disrupted. Revenue protected.
Cheapest intervention first. Always. Escalate only if evidence demands it.
The reasoning model

Every decision is explained. Not just delivered.

BAT · LFP · B · Week 47 · Fleet asset · Indian field conditions
Believe
This battery will breach the floor threshold within 6 weeks.
Based on
14 consecutive weeks of CUSUM drift below baseline. PELT structural break confirmed at Week 43. Conformal forecast: 82.4% probability of breach before Week 53. Corroboration: 4 of 7 signal channels agree.
Unless
Charger is replaced this week and high-rate charging behaviour changes. Cell spread historically recovers 40% within 3 weeks of charger correction in this fleet cohort.
Because
Cell spread acceleration has preceded floor breach by 11 weeks in 14 of 17 comparable batteries in the Pack3001 cohort. This is the pattern. Not an anomaly.
This is not a report.
This is the agent working.
Human-in-loop approval · Physics-grounded · Auditable at every step

Numbers and outputs shown are representative of enerlytik platform outputs running on Indian LFP field data. Individual battery outcomes vary by pack chemistry, operating conditions, fleet profile, and data availability. Percentages reflect attribution model outputs and should not be construed as legal determinations of liability. AI-generated recommendations should be reviewed by a qualified engineer or credit officer before action is taken. enerlytik does not guarantee specific financial, operational, or regulatory outcomes.

If this is your problem, let's talk.

gaurav@enerlytik.com · Gaurav Mehra · Co-Founder
Talk to us →