The 7-Phase AI Pipeline That Reads Your Utility Bills
Every emission figure from a utility bill traced back to the line on the source document, with the factor version and unit conversion recorded, and uncertain reads sent to a reviewer instead of guessed. A phase-by-phase walkthrough of the pipeline and where each phase breaks.
The result of a good utility bill pipeline is a reported tonne you can trace back to the line on the bill, with the factor version and unit conversion recorded, and uncertain reads flagged for a person instead of guessed. We built ours to do that at scale. And the thing that surprises us most isn't what the AI gets right, it's how creatively wrong documents can be.
A water bill from a regional council where someone scrawled the meter reading in pen over the printed estimate. A 14-page gas invoice where consumption data sits on page 9, sandwiched between regulatory notices nobody reads. An electricity bill photographed at an angle on a desk covered in coffee rings. These aren't edge cases. This is what real utility bill data extraction for ESG reporting actually looks like.
The reason we built a 7-phase pipeline - instead of one model that does everything - is because no single AI agent handles all of this well. An agent that's great at reading damaged scans is mediocre at unit conversion. One that nails emission factor lookups can't tell a water bill from a waste manifest. So we split the problem into seven discrete stages, each handled by a specialised agent, with explicit handoffs and validation gates between them.
This post is a technical walkthrough of that pipeline. What each phase does, what goes wrong, and the engineering decisions behind it. If you're evaluating AI document processing for carbon accounting, this is the detail most vendors won't show you.
Why One Model Isn't Enough
The temptation with large language models is to throw everything at a single prompt. Upload a bill, ask for structured JSON, get your answer. We tried this. It works most of the time.
The rest of the time will wreck your NGER submission.
The failure modes are specific and predictable. A single-model approach hallucinates numbers when it can't read them clearly - filling in a plausible-looking consumption figure instead of flagging uncertainty. It confuses billing period dates when they appear in multiple places on the document. It silently converts units wrong because it doesn't know that this particular gas retailer reports in megajoules while that one uses cubic metres.
Carbon reporting is high-stakes structured extraction, where even a modest error rate compounds across thousands of documents. When the CER audits your NGER submission, a plausible-looking number that didn't come from the source document is worse than a blank cell - because nobody catches it until it's a compliance problem.
Multi-agent architectures solve this by making each agent's job narrow enough that failure is detectable. If the classification agent is uncertain whether a document is an electricity bill or a gas invoice, it says so. It doesn't guess and pass garbage downstream. That explicit uncertainty is the whole point.
Phase 1: Document Classification
Every document enters the pipeline with no assumptions. We don't trust filenames (people name files "scan_003.pdf" and move on). We don't trust folder structures. The classification agent examines the document itself and determines what it is: electricity bill, gas invoice, water statement, fuel receipt, waste manifest, or something we don't handle.
This matters more than it sounds. An electricity bill feeds into Scope 2 calculations using grid emission factors. A natural gas invoice is Scope 1 - direct combustion. A diesel fuel receipt is also Scope 1 but uses completely different NGA emission factors. Classify wrong, and every downstream calculation is wrong even if the data extraction was perfect.
The classifier runs on document layout features, not just keywords. A document that says "Energy" in the header could be electricity or gas. The classifier looks at the presence of kWh vs MJ vs cubic metres, retailer-specific formatting patterns, and the relationship between charge types and units.
Where it struggles: combined utility bills. Some retailers - particularly for small businesses - issue a single document covering electricity and gas. The classifier needs to flag these as multi-commodity and route them so both commodity types get extracted separately.
Phase 2: Vision-to-Text via Multimodal AI
This is where the LLM earns its keep. Our multimodal AI vision reads the document image - scanned PDF, phone photo, born-digital invoice - and produces a structured text representation of everything on the page.
The distinction from traditional OCR matters. Template-based OCR systems need someone to define where on the page each field appears. "Total consumption is in row 14, column B." That works until a retailer changes its bill layout. Or until you receive an invoice from a retailer you've never seen. Then someone has to build and test a new template.
The model reads documents the way a human does. It understands that a number appearing after the words "Total Usage" and before "kWh" is a consumption figure, regardless of where on the page those elements sit.
We feed the model high-resolution document images and get back a structured representation that preserves the spatial relationships between fields. Tables, headers, footnotes, regulatory notices - all of it comes through with positional context intact.
Where it struggles: heavily degraded scans. A bill that went through a fax machine in 2019 and got scanned at 72 DPI is genuinely hard. Handwritten annotations over printed text cause problems too - the model sometimes reads the handwritten figure and sometimes the printed one, and it doesn't always tell you which. We've added pre-processing steps (contrast enhancement, rotation correction, resolution upscaling) that help, but we won't pretend we've solved the bottom 5% of document quality. Some bills need a human to look at them. That's just honest.
Phase 3: Structured Data Extraction
The vision-to-text output is rich but messy. It contains everything on the document - marketing messages, payment terms, tariff structures, regulatory fine print. The extraction agent's job is to pull out exactly the fields that matter for emissions calculations and nothing else.
Those fields are specific:
- Consumption quantity (the number)
- Unit of measurement (kWh, MJ, kL, litres, cubic metres, tonnes)
- Billing period start and end dates
- Meter number or NMI (National Metering Identifier)
- Site address
- Supplier name
- Whether the reading is actual or estimated
That last one - actual vs estimated - trips people up constantly. Estimated reads are common when a meter reader can't access a site. The bill still shows a consumption figure, but it's the retailer's best guess. If you're calculating Scope 2 emissions from electricity bills, using an estimated read without flagging it means your emission figure might get revised when the actual read comes through. Check the actual/estimated marker on every bill before relying on the figure.
The extraction step outputs structured data with each field and its value, and Carbonly keeps the verbatim source line text as evidence. That source reference is critical for Phase 7 - the audit trail.
Where it struggles: multi-page bills where consumption data and billing period dates appear on different pages. The agent needs to correlate information across pages, and sometimes the billing period on page 1 refers to a different period than the consumption breakdown on page 4. We handle this by having the agent process the full document and cross-reference internal consistency, but edge cases still exist - particularly with large commercial accounts that have multiple meters on a single bill.
Phase 4: Cross-Validation
Here's where many AI extraction tools stop. They get the data out and call it done. We don't stop there, because extraction without validation is just faster data entry with the same error rate.
Validation checks the extracted data against multiple signals. Is the consumption figure within a reasonable range for this type of site? A 200-square-metre office in Parramatta using 2,400,000 kWh in a quarter is obviously wrong - that's a smelter, not an office. Does the billing period align with previously processed bills from the same site? Is there a gap or overlap? Does the unit of measurement match what this retailer typically uses?
In Carbonly, low-confidence matches and units that cannot be converted go to a reviewer rather than silently passing through, and deterministic anomaly checks run daily at 6am and after each document batch. The reviewer sees the original document, the extracted data, and the reason for the flag.
Calibrating the sensitivity is a balance. Too loose and errors get through. Too tight and you're drowning reviewers in false positives.
Phase 5: Unit Normalisation
This phase is unglamorous and absolutely critical. Every consumption figure needs to be in consistent units before emission factors get applied. Sounds simple. It isn't.
Electricity is usually in kWh, but some retailers report in MWh for large commercial accounts. Gas comes in megajoules, gigajoules, or cubic metres depending on the retailer and the state. Water appears in kilolitres or litres. Waste shows up in tonnes or cubic metres (and those aren't interchangeable without density assumptions that change by waste stream).
Normalisation converts everything to the base units that NGA emission factors expect. For electricity, that's kWh. For natural gas, that's gigajoules. For liquid fuels, that's kilolitres. Where a conversion would need a figure the document does not state, such as a density, Carbonly sends the line to a reviewer instead of guessing.
The gotcha people miss: natural gas heating values. When a gas bill shows consumption in cubic metres, you can't just convert to gigajoules using a fixed multiplier. The heating value of natural gas varies by distribution zone. In Western Australia, it's different from Victoria, which is different from Queensland. Use the heating value shown on the bill for that distribution zone rather than a fixed multiplier.
We also handle the billing period normalisation here. If your reporting period is the financial year (July to June, as required for NGER) but your electricity bill runs from the 15th of one month to the 14th of the next, those periods don't align. The usual fix is to pro-rate consumption to the reporting period using daily consumption rates. It's not perfect - consumption isn't actually linear across a billing period - but it's a lot better than ignoring the mismatch.
Phase 6: Emission Factor Application and Calculation
Now we get to why all the previous phases exist. The calculation step takes normalised consumption data and applies the correct emission factors from the NGA Factors 2025 workbook to produce tonnes CO2-e.
For Scope 2 electricity, this means state-based grid emission factors. And they vary significantly. Victoria's grid factor is 0.78 kg CO2-e/kWh. South Australia's is 0.22. Tasmania's is 0.20. If you use the national average of 0.62 for everything - a common shortcut - you'll overstate SA emissions by nearly threefold and understate Victorian emissions by about 20%.
The correct factor depends on the site's location and the reporting year. Carbonly stores the emission factor version on each record, because NGA factors change annually. The 2025 factors show 2-3% reductions in grid intensity across most states compared to 2024-25 - NSW dropped from 0.66 to 0.64, Queensland from 0.71 to 0.67.
For NGER reporters, there's an additional wrinkle. NGER uses AR5 global warming potentials from the IPCC Fifth Assessment Report. AASB S2 requires AR6 values. The numerical difference is small for CO2 but material for methane (AR5: 28; AR6: 29.8 for fossil methane and 27.0 for non-fossil) and some refrigerant gases. If you're reporting under both frameworks - and every NGER reporter pulled into ASRS Group 2 is - you potentially need both sets of calculations from the same underlying data.
Carbonly writes AR5 values, the set NGER uses. An AASB S2 figure that needs AR6 has to be recalculated separately and labelled clearly so the right numbers go into the right submission. This is one of those details that sounds minor until you're explaining to an auditor why your ASRS Scope 1 figure doesn't exactly match your NGER figure.
Where we're still working: Scope 3 calculations from supplier invoices. The emission factors for purchased goods and services depend on the supplier's own emissions intensity, which varies wildly and is often unavailable. Spend-based emission factors can fill the gap, but those are averages across entire industry sectors. They're a placeholder, not an answer, which is why Carbonly floors spend-based Scope 3 to low data quality. We're honest about that limitation because overstating the accuracy of Scope 3 numbers is exactly the kind of claim that gets companies in trouble with the ACCC.
Phase 7: Audit Trail Generation
Every number produced by the pipeline needs to be traceable back to its source. Not next week when an auditor asks. Right now. Automatically.
The audit trail should link every emission figure to:
- The final calculated emission (e.g., 847 tonnes CO2-e)
- The emission factor used and its source (e.g., NGA Factors 2025, Table 1, Victorian grid, 0.78 kg CO2-e/kWh)
- The normalised consumption that was multiplied (e.g., 1,085,897 kWh)
- Any unit conversions applied (e.g., 1,085.9 MWh converted to kWh)
- The raw extracted value from the document (e.g., "Total Usage: 1,085.9 MWh")
- The original source document itself, stored as an immutable reference
That chain - from reported number back to the text on the source document - is what lets the Clean Energy Regulator or an auditor follow a figure in your NGER submission. It's what ASRS assurance providers will look for under ASSA 5010.
The ANAO found that 72% of 545 NGER reports they examined contained errors. Seventy-two percent. Errors of that kind often trace back to a gap somewhere between the source document and the reported number. The audit trail doesn't prevent errors in the AI pipeline (that's what Phases 4 and 5 are for). What it does is make every error findable and fixable, instead of buried in a spreadsheet that nobody can reconstruct.
What This Pipeline Can't Do
We'd rather be upfront about the limitations than have you discover them after you've committed to a tool.
Handwritten documents - entirely handwritten utility records (still common in some regional water authorities) are unreliable. The pipeline will attempt extraction but flags everything as low confidence. You'll need human review.
Non-English bills - if you've got supplier invoices in Mandarin or Japanese for Scope 3 supply chain reporting, expect lower extraction accuracy and plan for more human review.
Consolidated corporate invoices - some energy retailers issue a single invoice covering 50+ sites with a summary table and individual site details across dozens of pages. We handle these, but processing time is longer and the validation phase generates more flags because there are more opportunities for cross-page correlation errors.
Historical documents - bills from before 2015 tend to have different formats, lower scan quality, and sometimes use emission factors that have since been updated. The pipeline extracts data but can't retrospectively apply the correct historical NGA factors without manual configuration.
We think radical honesty about what AI can and can't do is more useful than a marketing page claiming 99.9% accuracy. The 99.9% number that some vendors quote typically comes from controlled benchmark conditions with clean, high-resolution, well-formatted documents. The real world includes coffee-stained photocopies and 14-page gas invoices. Accuracy across the full range of real document quality is lower, which is why review and validation matter. It isn't magic.
Why the Architecture Matters More Than the Model
The current vision model will be replaced by a better one. Probably within a year. The specific LLM sitting behind our vision-to-text phase is a component, not the product.
What matters is the architecture: the separation of concerns between agents, the validation gates between phases, the explicit uncertainty handling, and the audit trail that connects reported numbers to source documents. That architecture works regardless of which foundation model powers Phase 2. When we upgrade the underlying model - and we will - the extraction accuracy improves but the pipeline structure stays the same.
This is why we built Carbonly as a multi-agent system rather than a wrapper around a single API call. A wrapper breaks when the API changes. A pipeline with discrete, testable phases breaks in predictable, fixable ways.
If you're evaluating AI-powered carbon accounting tools, ask the vendor what happens when their model hallucinates a number. If the answer is "it doesn't" or "we have 99.9% accuracy," keep looking. The right answer is: "we detect it here, flag it for review, and maintain a full audit trail so you can verify any figure back to its source document."
That's not a feature. It's the minimum viable requirement for regulatory-grade emissions reporting. To see that chain on one of your own bills, email hello@carbonly.ai.
FAQ
Can software read utility bills and calculate Scope 2 emissions automatically in Australia? Yes. Carbonly reads electricity and gas bills and takes the consumption, unit, billing period and site. For electricity it applies the NGA grid factor for that state. The factor version is stored on each record, and the line text from the bill is kept as evidence. Uncertain reads go to a reviewer. Carbonly calculates on AR5, the set NGER uses.
How do I stop keying electricity and gas bills into a spreadsheet for carbon reporting? Move the reading of the bill to software and keep your people for review. The fields you key today are few: consumption, unit, billing period, meter number or NMI, site and supplier. Software that reads those fields from each bill, flags estimated reads and holds the bill beside the figure removes the typing. Your team then checks the flagged bills.
What happens when a bill is a poor scan or covers several meters? It should go to a person for checking. A badly degraded scan, or a bill with handwriting over the print, can be misread. Bills covering several meters need the usage matched to the right meter and billing period across pages. In Carbonly, low-confidence reads go to a reviewer, who sees the original document beside the extracted figures.
Should I use an estimated meter read in my emissions figures? You can use it, but mark it as estimated. An estimated read is the retailer's best guess when the meter could not be accessed. The figure may be revised when the actual read comes through, which changes the emissions calculated from it. Check the actual or estimated marker on every bill, and record it beside the consumption figure.
Why does my gas bill in cubic metres need a heating value to convert to gigajoules? Because the energy in a cubic metre of natural gas varies by distribution zone, so a fixed multiplier can misstate it. NGA emission factors for natural gas are applied to gigajoules. Use the heating value shown on the bill for that distribution zone to convert cubic metres to gigajoules, then apply the factor. Record the conversion beside the figure.
How do I handle a bill whose billing period crosses the end of the reporting year? Pro-rate the consumption to the reporting period using a daily rate. NGER reports on the financial year from July to June, while a bill may run from mid-month to mid-month. Divide the usage by the days in the billing period, then allocate the days that fall in each year. Usage is rarely even across a period, so note the method you used.
Related Reading:
- AI Document Processing for Carbon Accounting - the business case for automated extraction
- How to Calculate Scope 2 Emissions from Australian Electricity Bills - the formula and NGA factors this pipeline applies
- NGER Compliance Automation - how automated pipelines reduce compliance risk
- NGER Reporting Thresholds 2026 - who needs to report and what the CER expects
- How One AI System Reads 200+ Utility Bill Formats
- Carbon Accounting for Construction Companies in Australia - where document volume breaks manual entry first