How AI Energy Plans Calculates Comparisons and Savings Estimates

1. Electricity-Rate and Plan Data Sources

Every plan referenced in a comparison traces back to a specific, identifiable source document. For Texas, this means the EFL associated with that plan, which discloses the base charge, energy charge structure, any Time-of-Use windows, contract length, cancellation fee, and estimated pricing at standard usage tiers (500, 1,000, and 2,000 kWh). Where a provider publishes updated terms, promotional credits, or bill-credit structures, those are captured from the provider's own documentation rather than inferred.

Because EFLs can use inconsistent formatting from one provider to the next: some list rates as flat tables, others embed Time-of-Use tiers or seasonal rate matrices in narrative text: extracting this data reliably requires a structured parsing process, described further in the How AI or Automated Analysis Is Used section below.

2. Update Frequency

Retail electricity plans in ERCOT's deregulated market can change frequently: providers introduce new plans, retire old ones, and adjust promotional pricing on their own schedules. AI Energy Plans checks for updated plan and rate information on a recurring basis, and each plan record reflects the date it was last verified against its source EFL. Because market conditions and provider offerings shift, we recommend cross-checking any plan you are seriously considering against the provider's current, published EFL before signing a contract.

3. Household Usage Assumptions

Generic usage tiers are a useful starting point for comparing plans at a glance, but they don't capture a specific household's actual behavior: home size, occupancy, appliance mix, EV charging, and HVAC habits all affect real consumption. Wherever possible, the platform is designed to move a user from generic tier-based comparisons toward a personalized estimate based on their own monthly usage history, since that materially improves the accuracy of any bill or savings projection.

4. Time-of-Use Calculations

TOU plans price electricity differently depending on the hour of day and, in some cases, the season. To estimate a household's cost under such a plan, we need to know not just total monthly usage but roughly when that usage occurs. Where interval (hourly or 15-minute) usage data is available: for example, from a smart meter export a user provides: that data is used directly. Where it isn't, the estimate relies on typical Texas residential load-shape assumptions (e.g., higher afternoon and evening usage in summer). This is one of the areas where estimates for a specific household can diverge most from an average household, particularly for homes with unusual schedules, EV charging outside typical hours, or battery-arbitrage dispatch.

5. Battery Round-Trip Efficiency

No battery stores and returns energy with perfect efficiency. Charging and discharging losses occur in the battery cells, the inverter, and other system components. AI Energy Plans applies a round-trip efficiency assumption consistent with published residential battery specifications when modeling energy-arbitrage strategies (charging during low-price periods and discharging during high-price periods) so that projected savings reflect real losses rather than an idealized exchange rate.

6. Battery Degradation and Dispatch Assumptions

A battery's usable capacity declines gradually across its lifespan due to charge-cycle wear. Ignoring this would overstate long-term savings, so multi-year projections apply a degradation curve rather than assuming day-one capacity holds constant for a decade or more. Dispatch logic: the rules governing when the battery charges from the grid or solar, and when it discharges to serve the home: is modeled based on the scenario a user selects, such as prioritizing TOU arbitrage versus prioritizing backup reserve. Real-world dispatch behavior depends on the specific battery, inverter, and control software a household actually installs, which may behave somewhat differently than the modeled scenario.

7. Solar Production Assumptions

Estimating how much electricity a solar array will produce requires combining the physical characteristics of the system (its size, tilt, azimuth, and any shading) with the amount of sunlight the site historically receives. AI Energy Plans applies standard solar-resource assumptions for the relevant Texas location along with commonly used system-loss factors: covering inverter conversion losses, wiring losses, panel soiling, and temperature-related derating: to translate expected sunlight into an estimated kWh output. Two homes with identically sized systems in different parts of Texas, or with different roof orientations, can have meaningfully different actual production.

8. Weather Inputs

Weather affects two parts of the model: how much sunlight a solar system will receive, and how much electricity a household is likely to use for heating and cooling in a given season. Both use historical climate data as a proxy for typical conditions rather than attempting to predict specific future weather. Actual weather in any given year: an unusually hot Texas summer, a mild winter, or a cloudy stretch: will cause real solar output and real usage to differ from the modeled baseline.

9. Taxes, TDU/Delivery Fees, and Other Charges

A Texas electricity bill is really two charges combined: the retail energy charge (set by the competitive provider you choose) and the TDU delivery charge (set by the regulated utility that owns the physical wires in your area, and identical for every provider serving that territory). Because delivery charges differ by TDU and are periodically adjusted, estimates use the delivery rate applicable to the service area being analyzed. Any other recurring fees disclosed in a plan's EFL: such as minimum usage fees or monthly service charges: are also incorporated where applicable.

10. Payback Methodology

Payback calculations combine several of the inputs described above: solar production estimates, battery dispatch and degradation assumptions, the household's electricity plan and rate structure, and the applicable TDU delivery charges. The result is a projected annual savings figure, which is compared against the net system cost to estimate a simple payback period. This calculation does not attempt to model complex financial factors like the time value of money, financing interest, or future changes in electricity rates or tax incentives unless a user explicitly provides that information.

11. Limitations and Uncertainty

12. How AI or Automated Analysis Is Used

Two different jobs are involved in producing a plan comparison, and it's worth distinguishing them clearly:

In short: AI helps us read and organize rate documents faster and at scale; it does not invent the numbers used in your comparison. This distinction matters for auditability: anyone reviewing an estimate on this site can trace it back to a specific EFL, a specific usage assumption, and a specific formula, rather than an opaque AI-generated guess.