AI Battery Copilot & physics informed ML models

AI and ML for batteries at scale

  • Empirical approach
  • Expert reviewed
  • Evidence backed

Ensuring project proforma and profitability

🟣 Live Copilot Demo

AI Battery Copilot

Talk instead of sifting through dashboards

💬

Issue commands to AI

Use private agent or add to your Slack, Microsoft Teams or email

📊

Connects to Telemetry

Realtime BMS & PCS data through APIs

💹

Connects to Markets

Live energy prices across markets via APIs

⚙️

Runs ML Models

Physics-informed Machine Learning Models based on cell cycling and research

📚

Reads Knowledge Base

AI follows proprietary knowledge base before responding

👤

Keeps Human-in-the-Loop

All responses are backed by data evidence and are escalated if below confidence level

🧠

Engages Strategic Consultants

1-1 call with energy and AI consultants for critical decisions and architectural changes

Battery Copilot Online · <1 min response
Domain expert
AM

Show me SoH trends for Rack 12. Is it degrading faster than expected?

VP-Assets · 09:02
Copilot

Rack 12 is at 94.2% SoH after 847 cycles — 1.8% ahead of warranty curve. Here's the trend:

Cycles vs State of Health
0 500 1000 100% 90% warranty 94.2%
Battery Copilot · 09:02
AM

Good. What's the revenue at risk if we take it offline this week for maintenance?

CBO · 09:03
Copilot

Revenue at risk ≈ $62k this week based on forward prices. Recommend a 4-hour maintenance window tonight during off-peak. Alert sent to O&M team.

Battery Copilot · 09:03
CF

How are we tracking against proforma? Show me DSCR and equity IRR projections.

CFO · 09:04
Copilot

DSCR at 1.32x (above 1.2x floor). Equity IRR tracking 12.4% vs 12.0% proforma target. Degradation 0.8% better than modeled. Detailed variance report sent to your inbox.

Battery Copilot · 09:04

AI-assisted battery asset management

🛡️ Reduce unplanned outages

Early warnings for safety and performance risks that BMS or PCS miss.

🔋 Unlock full usable capacity

Accurate SoC calibration to increase revenue per operational MWh.

💸 Evidence backed warranty claims

AI-assisted warranty reports with clear operational evidence and root causes.

📉 Degradation cost control

SoH diagnostics to manage degradation cost and protect IRR.

⏱️ Faster time to revenue

In-silico simulations to right-size projects before buying hardware.

⚡ Better market price capture

Forecast and dispatch to increase revenue realisation from price spikes.

⚖️ Smarter scheduling coordination

AI-assisted schedules aligned to asset constraints and revenue goals.

How We Lift Battery Asset Profitability

Physics-informed ML models and AI automation tuned for BESS economics.

Physics-Informed ML

Physics-Informed ML Models

Example conversation

Asset Manager

What's the projected SoH at month 36 for our LFP cells?

Battery Copilot

91.2% SoH expected at month 36 based on current cycling profile. That's 2.3% above warranty floor—within safe margin.

0 18 36 mo 100% 85% warranty 91.2%

Predict degradation, downtime and usable capacity with physics-based models that speak in CFO metrics.

  • Approach: Blend electrochemical twins with telemetry to surface availability risks before they become forced outages.
  • Pre-trained for your chemistry: Like ChatGPT, we start with chemistry-specific pre-training and fine-tune on your cycles.
  • Impact from day one: Models arrive tuned, so KPI lift is visible on first ingest.
Free SoH model demo

Simulations Lab

Simulated Scenarios

Example conversation

CFO

Compare 4h vs 6h duration for Texas ERCOT market mix.

Battery Copilot

6h duration lifts IRR by 140 bps to 12.8% with ERCOT's peak spread. Payback improves 8 months vs 4h config.

4h duration 6h duration 11.4% 12.8% +140 bps

Run what-if studies before pouring concrete so lenders see the upside, and you see the downside guardrails.

  • What-if analysis: Rapidly test financing and operating decisions.
  • Profile-aware models: Simulate chemistries, configurations and dispatch profiles to see impact on IRR/DSCR.
  • Investment memos: Export lender-ready guardrails for capital committees.

Revenue Stack

Revenue Optimization

Example conversation

Operations Lead

Should we discharge now or wait for the evening ramp?

Battery Copilot

Hold. 78% probability of $145/MWh spike at 18:00 based on weather + outage forecast. Current offer is $62/MWh.

12:00 15:00 18:00 $150 $50 $62 $145

Turn telemetry and market feeds into dispatch calls aligned with your project proforma instead of dashboards to stare at.

  • Stochastic fundamental modeling: Randomized weather, outages and price paths capture upside spikes.
  • Multi-market dispatch: Allocate charge/discharge across wholesale, ancillary and capacity markets with confident overbooking.
  • Portfolio optimization: Pair storage with renewables to smooth risk and boost IRR.
  • Right-sizing calculator: Confirm duration and power sizing before hardware deposits.

Enterprise BESS teams we equip

Purpose-built playbooks for the people accountable for revenue realisation, DSCR and availability.

Developers & EPC

Accelerate time-to-revenue

  • Development margin / MW: Quantify usable capacity and losses across commissioning.
  • Validated design envelope: Compare model simulations vs. as-built telemetry.
  • Time to revenue: Close punch lists faster with evidence packs for OEMs and lenders.

Outputs: commissioning copilot, test automation, digital twin for pre-COD bids.

Asset owners & IPPs

Protect equity IRR

  • Revenue realisation: Dispatch recommendations across wholesale, ancillary and capacity markets.
  • Proforma tracking: Monitor deviations from project proforma and run sensitivity analysis to protect revenue targets.
  • Equity IRR & DSCR: Scenario planning with IRR, DSCR and downside risk views.
  • Residual value: Long-term degradation tracking for refinancing and re-powering.

Outputs: portfolio cockpit, investment committee briefings, warranty dossiers.

Operators & O&M partners

Deliver availability targets

  • Availability & forced outage rate: Streaming anomaly detection and automated escalation.
  • O&M cost / MW-year: Predictive maintenance that prioritises the highest revenue-at-risk racks.
  • Warranty cost / MWh: AI-generated root cause analysis with recommended work orders.

Outputs: shift briefings, alert routing to email/Slack/Teams, reliability notebooks.

🟣 Live Copilot Demo

BESS Market Explorer

Project tracker for BESS deployments across our key markets

Operational
Under Construction
Announced
0 Total Projects
0 Operational
0 GWh Capacity

Apply AI & ML to your profitability

No generic demos—we listen to your KPIs first.

💌 Schedule a Meeting

Email: hello@amperical.ai