Australian Multi-Site Utility Bill PDF to Excel

An Australian commercial property manager running 12 sites across NSW and VIC, a franchise head office processing bills for 40 owned-and-operated locations, an energy procurement consultant preparing a tariff review, or a bookkeeper servicing portfolio property clients — every one of them faces the same recurring monthly job. PDF bills arrive by email from AGL on some sites, Origin Energy on others, EnergyAustralia or Red Energy or Alinta Energy elsewhere, gas bills under separate account numbers, and water bills from Sydney Water, Yarra Valley Water, South East Water, SA Water, Water Corporation, or whichever authority covers each supply address. The job is to consolidate them into one per-site spreadsheet that holds up for cost reporting, tenant recharge, and ESG disclosure.

The Australian multi-site utility bill PDF to Excel workflow rests on three structural facts. First, every electricity bill in Australia carries a National Metering Identifier (NMI) tied to the connection point, not the retailer — so when a site moves from AGL to Origin, the NMI stays the same and remains the per-site join key across the portfolio. Second, every gas bill carries a Meter Installation Reference Number (MIRN) that does the same job for gas. Third, water bills carry a property reference number issued by the water authority, which plays the same role for water. These three identifiers are how a portfolio's bills consolidate cleanly across providers, periods, and historic retailer switches.

The destination is a single per-site, per-period row with the bill's tariff structure broken out into columns the spreadsheet actually needs: supply charge separated from usage charge, peak / off-peak / shoulder kWh on their own columns where time-of-use applies, demand kVA as a separate column for medium-business sites, controlled-load kWh for separately metered loads like off-peak hot water, GST on its own line, and the retailer or authority name retained so cross-vendor comparison is direct. One row per site per billing period, ready for a pivot on site code, year-on-year comparison, tenant recharge calculation, or kWh-to-tCO2-e conversion for Scope 2 reporting.

This article is the Australian regional companion to the generic utility bill PDF to Excel workflow — the multi-provider AU reality with vendor and tariff specifics that a one-size-fits-all approach cannot carry. The Big 5 retailers each lay their bills out differently, the metro water utilities sit invisibly in most utility-bill writing despite being a monthly reality for property managers, and the AU tariff vocabulary (demand tariff, controlled load, regulated versus market offer, embedded networks) is what the spreadsheet columns are built from.

---

The Big 5 Retailers and What Their Bills Look Like for Extraction

AGL, Origin Energy, and EnergyAustralia are the Big 3 retailers a portfolio will encounter most often, with Red Energy (Snowy Hydro-owned) and Alinta Energy completing the practical Big 5 for commercial bills. The same retailer often covers electricity and gas at the same site, but a portfolio that has changed hands, switched plans, or grown through acquisition is rarely on one retailer for everything. A property manager looking at last month's mailbox sees three or four logos at minimum, and the extraction has to handle them as one job rather than five.

AGL. Business and commercial bills carry the NMI prominently in the account summary block at the top of page one, alongside the supply address and the named tariff plan (AGL Business Saver, AGL Energy Plus, and similar plan-specific names). Where time-of-use applies, the peak / off-peak / shoulder breakdown sits inside a usage detail table with kWh and rate per period. Medium-business sites with a demand component show a separate kVA or kW max-demand line in the charges section. AGL's business portal also offers NEM 12 download as a separate function, which is a usage-analytics input rather than a substitute for the bill. For an AGL bill PDF to Excel pass, the prompt has to name the NMI, supply charge, usage charge, peak / off-peak / shoulder kWh and rates, demand kVA where present, and GST.

Origin Energy. NMI placement is in the supply details panel rather than the account summary header, and the tariff naming is plan-specific. Origin's commercial bill layout breaks supply charge and usage charge cleanly into separate line items, with controlled-load lines listed alongside where applicable. The AER reference price comparator that appears on residential Origin bills is absent on commercial bills, which means an Origin Energy bill to Excel extraction reaches for the same field list the AGL pass does — only the layout positions move.

EnergyAustralia. Bill layout puts the NMI alongside the supply address rather than in a header block, and the time-of-use windows are stated explicitly on the bill — peak hours, shoulder hours, and off-peak hours per region — which makes the period rates easy to read into rate columns when they are needed. Demand tariff sites carry a separate demand summary block, and EnergyAustralia is the retailer most commonly seen on multi-site business arrangements where one plan covers many connection points. Every connection point still has its own NMI, so an EnergyAustralia bill PDF to Excel output lands one row per NMI even when the cover account is shared — row count tracks the connection points, not the account.

Red Energy. Commonly selected on Victorian and NSW sites. The NMI appears in the supply summary alongside the connection address, and the bill layout is one of the cleaner ones for extraction — usage charges are itemised by period clearly, and the supply charge is on its own line above the usage block. The textbook-bill structure means a Red Energy bill to Excel pass tends to inherit whatever prompt the rest of the portfolio is using without modification.

Alinta Energy. Strongest presence in WA and SA gas, and Alinta's commercial gas bills carry the MIRN in the place where NMI sits on electricity bills. Alinta also retails electricity in eastern states, with a layout distinct from the eastern-states retailers that is worth noting in the prompt. Because the same retailer covers both fuels at many sites, the Alinta Energy bill PDF to Excel output uses a vendor-type column to distinguish electricity rows from gas rows — the rest of the field list stays shared.

Gas bills from any of these retailers follow the same structural shape with three substitutions. The MIRN replaces the NMI as the per-site join key; consumption is in megajoules (MJ) rather than kWh; meter read dates and the daily standing charge sit where electricity's supply charge sits. Otherwise the supply-charge / usage-charge / GST structure is parallel, so a commercial gas bill to Excel pass shares the prompt with the electricity pass — the prompt asks for MJ where the document is gas, kWh where it is electricity, and the AI handles the per-document field mapping.

The reason a multi-provider portfolio cannot be served by a fixed-template extraction tool — the kind that expects one trained layout per vendor — falls out of the per-retailer notes above. Layout placement varies, plan names vary, header block structure varies, and a portfolio that adds a new retailer next quarter would need a new template. The extraction prompt has to describe the fields by name — NMI, supply address, billing period, supply charge, peak kWh, demand kVA — and let the AI find each field in whichever layout the bill arrives in.

---

The Water Utility Layer: State-by-State, Property Reference as the Per-Site Key

Water in Australia is not deregulated. There is no choice of retailer, and a property manager does not switch water providers the way they might switch from EnergyAustralia to Red Energy. Each region has a single water authority, and the bill for any given supply address comes from whichever authority covers that postcode. The corollary for extraction: the water layer of a portfolio's monthly bills is multi-authority by virtue of geographic spread, not multi-retailer by virtue of choice — but the consolidation problem is the same shape.

The metro and state authorities a multi-site portfolio actually encounters:

• NSW metro: Sydney Water for greater Sydney, the Illawarra, and the Blue Mountains; Hunter Water for Newcastle and the Lower Hunter.
• VIC metro: Yarra Valley Water for north and east Melbourne; South East Water for the south-east; Greater Western Water (formed from the merger of City West Water and Western Water) for the north and west.
• QLD metro: Urban Utilities for Brisbane and surrounds; Unitywater for the Sunshine Coast, Moreton Bay, and Noosa.
• SA: SA Water statewide.
• WA: Water Corporation statewide.
• TAS: TasWater statewide.
• NT: Power and Water Corporation.
• ACT: Icon Water.

The per-site join key on a water bill is the property reference number — each authority calls it slightly differently (property number, account number, supply address ID), but the function is identical to the role the NMI plays for electricity and the MIRN plays for gas. It is stable across billing periods, tied to the supply address rather than the customer, and printed on every bill. A Sydney Water Yarra Valley Water bill to Excel pass across a portfolio with NSW metro and VIC metro sites uses the property reference as the per-site key on each, and lets the vendor-type column distinguish the authority.

A commercial water bill carries components that have to land as separate columns rather than a single bill total. Water usage charge is per kilolitre (kL) and is sometimes tiered, with a higher rate above a threshold of consumption. The water service charge is the fixed daily or quarterly component covering the supply itself. Sewerage and wastewater service charges sit on their own lines, with the sewerage usage component sometimes calculated as a percentage of water usage rather than directly metered. Trade waste charges apply to commercial sites that discharge waste exceeding domestic-equivalent standards (food service, manufacturing, healthcare) and need preserving as their own line. Drainage charges appear on bills covering certain stormwater catchments. GST on water bills is line-by-line: water for residential use is GST-free in many cases, while commercial sites typically have GST applied on specific components rather than the bill total. Extraction has to preserve the line-level GST flag rather than collapse to a single total, or the GST column on the unified row will be wrong for water.

Water authorities issue bills quarterly more often than monthly, which is the structural difference that matters most for the unified spreadsheet. A site that has monthly electricity bills, monthly gas bills, and quarterly water bills lands four water rows per year against twelve electricity rows, and the period boundaries for water do not align with calendar months. The per-site, per-period row carries explicit billing-period-start and billing-period-end columns rather than assuming a calendar month, so cross-vendor analysis aligns by date range rather than by month name. A pivot on quarter, not on month, is the right view for any cross-utility comparison that includes water.

---

NMI, MIRN, and Supply Address: The Join Keys That Survive Multi-Provider Reality

The reason address strings fail as a join key across a multi-provider portfolio is that retailers store the same physical site differently. "Suite 4, 12 Smith Street" on the AGL bill becomes "12 Smith St, Suite 4" on Origin and "12 Smith Street Unit 4" on EnergyAustralia. A spreadsheet joined on address produces three rows where there should be one, and any cross-period comparison falls apart when the same site switches retailer mid-year. The per-site join key has to be something the retailer cannot change.

For electricity, that key is the NMI. According to the AEMO MSATS National Metering Identifier procedure, an NMI is a ten-character identifier; a one-digit checksum may be appended to form an eleven-character NMI used to validate manual data entry. It is tied to the connection point, not the retail account — a site that switches from AGL to Origin keeps the same NMI on every bill before and after the switch, which is exactly the property a portfolio's join key needs. It is unique per electricity supply point, printed on every electricity bill in Australia regardless of retailer, and not subject to reassignment the way a customer number or phone-style account ID might be. A property manager whose 12 sites move between retailers across a five-year horizon ends up with 12 stable NMIs, even as the retailer column on the spreadsheet churns.

For gas, the parallel key is the MIRN — a per-site identifier issued by the gas distributor rather than the retailer, printed on every gas bill, stable across retailer switches in the same way the NMI is. Gas distribution in Australia is regional: Australian Gas Networks (AGN) covers parts of SA, VIC, the NT, and Queensland; Jemena Gas covers parts of NSW; AusNet Services and Multinet cover parts of VIC; ATCO Gas covers WA. The MIRN format is set by the distributor and is shorter than an NMI, but it functions identically as the join key. A portfolio with electricity NMIs and gas MIRNs maintains both as separate columns on the unified row, with a vendor-type column flagging which one the row applies to.

Supply address still belongs in the spreadsheet, but as a human-readable column rather than the join. The pattern that holds up: store the join key (NMI, MIRN, or property reference) as one column, and store a normalised supply address — one canonical form decided once for the portfolio, applied consistently regardless of which retailer's string the bill carries — as a separate column. Cross-period and cross-retailer joins run on the join key; the address column exists for the human reading the spreadsheet and for any external system that has not adopted the join key yet.

This is the spine of a portfolio's NMI-keyed electricity spreadsheet, and the spine of utility bill extraction across an Australian property manager's full site list more broadly. A single prompt-driven extraction pass over the portfolio's mixed-vendor PDFs produces the unified per-site row, with the join keys naming each row and the tariff fields broken out per the spreadsheet shape. AI utility bill extraction for Australian portfolios is the same one-job pipeline applied at portfolio scale: upload the month's bill PDFs as one batch, prompt the AI for the field list the spreadsheet needs, download the structured output. The product handles up to 6,000 files per batch and individual PDFs up to 5,000 pages, which matters when a quarter-end bill set is large or when a single retailer aggregates many sites onto one multi-page PDF. A goal-oriented prompt — "I'm extracting monthly utility charges across our sites for cost reporting, one row per site per billing period, with NMI as the join key, supply charge separated from usage charge, peak / off-peak / shoulder kWh broken out, demand kVA where present, controlled-load kWh where present, and GST on its own line" — runs across whichever retailer's layout the bill arrives in, and the AI maps the fields to the same row shape on every output.

The same multi-provider extraction pattern applies in other markets a portfolio might operate in: the Hong Kong multi-site utility bill extraction (regional sibling) covers the equivalent vendor and tariff vocabulary for HK — different retailers, different metering identifiers, same workflow shape with the join key, the per-site row, and the prompt-driven extraction operating identically.

---

Reading the AU Tariff Lines: What Each Component Means in Your Spreadsheet

The components on an AU commercial bill follow the vocabulary the AER and the retailers' own bill explainers use, and the unified row carries each component as its own column rather than a single bill total. Practitioners who have read enough bills will recognise most of this; setting it out in one place gives the prompt that drives the extraction a complete field list to work from.

Supply charge (fixed daily charge). Charged per day regardless of consumption, covering the network's standing service to the connection point. The bill multiplies the daily rate by the number of days in the billing period and shows the result as a discrete line. On the unified row, supply charge sits in its own column. Collapsing it into the usage charge breaks any cost-per-site analysis that wants to separate fixed cost from variable, and it breaks tenant recharge calculations where pass-through pricing only applies to one of them.

Usage charge (consumption charge). Per kWh on electricity, per MJ on gas, per kL on water. Where the tariff is flat, this lands as a single usage-total column. Where time-of-use applies, it splits across columns by period.

Peak / off-peak / shoulder. Time-of-use tariffs name three windows, with the time bands and rates differing by retailer and by network. The bill always names them — peak hours might be 2pm to 8pm on weekdays in one region, 7am to 11pm in another, with shoulder around the edges and off-peak overnight and weekend. The unified row carries three usage columns (peak kWh, off-peak kWh, shoulder kWh) and three rate columns where rates need preserving. Rate preservation matters most for tenant recharge, where the recharge agreement may pass peak rates through directly rather than averaging.

Demand tariff. Common on medium-business sites, typically those above approximately 100 MWh per year, though the threshold varies by network and retailer. The charge applies to the maximum demand reached during peak windows, measured in kVA or kW, multiplied by a per-kVA-per-month or per-kVA-per-day rate depending on the network. The demand line sits separately from the usage lines on the bill, and on the unified row it gets a demand column distinct from the usage columns. A site that crosses the demand threshold from one quarter to the next looks the same in usage terms but lands a new column value, which is a useful signal for procurement consultants reviewing tariff fit.

Controlled-load tariff. A separately metered load — most often off-peak hot water, slab heating, or pool pumps — billed at a different (usually lower) rate than the main meter. It lands on its own kWh column with its own rate, distinct from the main meter's usage. Sites with controlled-load arrangements typically have their NMI extended with a sub-identifier on the bill, but the join key for the spreadsheet stays the parent NMI.

Regulated offer versus market offer. In the deregulated states (NSW, VIC, QLD, SA), most commercial sites are on market offers — negotiated plans with rates outside the regulated cap. The Default Market Offer (DMO) is the regulated cap on standing offers in NSW, SA, and SE QLD; the Victorian Default Offer (VDO) is the equivalent in Victoria. The bill names which offer type applies, and on the unified row it sits as a metadata column. For an energy procurement consultant comparing portfolios this matters; for a bookkeeper running cost reporting it sits quietly until tariff renegotiation time.

GST line. GST is shown as a discrete line on the bill total, applied at 10% to taxable components. The unified row carries GST as its own column. The connection to BAS is direct: every utility bill flows GST into the activity statement, and a per-site, per-period row with GST broken out is exactly what a bookkeeper needs for preparing BAS in Xero for GST on utility expenses without further reshaping. Water bills handle GST line by line — usage and service charges may be GST-free or GST-applicable depending on the customer type, and the extraction has to preserve the line-level treatment rather than apply GST to the bill total only. A water row whose GST is calculated as 10% of the total when half the lines were GST-free overstates the input tax credit.

Embedded networks. Where a landlord, building owner, or shopping centre operator buys energy in bulk from a retailer and on-sells to occupants, the bill an embedded-network customer receives looks different from a retailer bill. Tariff labels resemble the retailer bill, but the supplier name is the embedded-network operator (often through a third-party billing service), and the tariff structure may not map exactly to a retailer plan because it reflects the on-sell margin and the metering arrangement. The extraction prompt handles embedded-network bills as a separate vendor type column value, with the same supply-charge / usage-charge / GST shape applied where the line items support it.

The extraction implication of all of this is that the prompt names every one of these fields explicitly. The prompt is the contract between the bill and the spreadsheet — supply charge, usage charge total, peak kWh, off-peak kWh, shoulder kWh, demand kVA, controlled-load kWh, GST line, plan name, retailer name. The AI finds each field in whichever retailer's layout the bill arrives in, and the unified row carries each component as a distinct column rather than a single total. That is what separates accounting-grade extraction — the kind that reconciles to the bank, flows GST cleanly, and supports tenant recharge — from the usage analytics a meter-data download provides.

---

NEM 12 Versus the Bill: When Each One Is the Right Answer

Anyone who has worked with an energy procurement consultant or used an AGL business portal will have heard the suggestion: "just download NEM 12". The advice is sound for one job and misleading for another, and conflating the two costs either accuracy or weeks of rework. NEM 12 is the Australian Energy Market Operator's standard CSV format for interval meter data — typically thirty-minute reads — for a single NMI. It is available for download from some retailer business portals (AGL is the most commonly cited) and from some metering coordinators directly.

What NEM 12 carries: usage in kWh per interval, the NMI, the meter identifier, and read quality flags. That is what it is designed to do, and it does it well.

What NEM 12 does not carry: no supply charge, no demand charge, no controlled-load rate, no GST, no network charges, no totals, no retailer name, no plan name, no billing period in the accounting sense. NEM 12 is meter data, not financial-document data. The two formats answer different questions, and only one of them is what the bill carries.

The use cases divide cleanly. Usage analytics and load-shape work is what NEM 12 is built for. Thirty-minute interval data is far more granular than the bill's peak / off-peak / shoulder summary, and a load-shape question — when does this site actually use energy, what is the demand profile, where is the optimisation potential, what would a different tariff structure cost — needs the interval reads. The bill's three-bucket summary cannot answer any of those questions because it has already aggregated the underlying data.

Accounting, cost-per-site reporting, tenant recharge, and BAS preparation is the other half, and the bill is the right input for all of it. The bill is what reconciles to the bank account, what GST flows through, and what tenants see when they get a recharge invoice. NEM 12 cannot do any of those jobs because the data needed for them — the dollar amounts, the GST split, the supply charge, the network charges — is not in NEM 12 at all.

The practical reality for most bookkeepers and property managers tilts the balance further toward the bill. Even where NEM 12 download is available, only large customers with portal access can pull it, and only for electricity — not gas, not water. Many bookkeepers servicing portfolio property clients receive bills as PDF email attachments and have no portal access on most of the underlying accounts. The bill is the only input that exists for the full portfolio.

A portfolio that needs both load-shape analysis and cost reporting often runs both inputs in parallel, with the bill extraction feeding the per-site spreadsheet for accounting and recharge, and NEM 12 feeding a separate analytics workflow for load shape and tariff optimisation. They are complementary, not substitutes — and confusing them is the most common reason a property manager spends a quarter trying to reconcile NEM 12 totals against bill totals and discovering the numbers were never meant to match.

---

The Unified Per-Site Row, in Two Variants

Every bill, regardless of retailer, vendor type (electricity, gas, water), or layout, lands in the same row shape so cross-vendor and cross-period analysis is direct rather than requiring per-vendor pivots. The two variants below cover the two jobs a portfolio is most often doing — analytical cost reporting and tenant cost allocation — and most portfolios maintain the analytical row as the master and derive the recharge view from it rather than running two separate extractions.

Cost-per-site analytical column shape. The full column list, each on its own line in the spreadsheet:

• Site code (internal identifier)
• Supply address (normalised)
• NMI / MIRN / property reference (the join key)
• Vendor type (electricity, gas, water)
• Retailer or authority name
• Plan or tariff name
• Billing period start
• Billing period end
• Days
• Supply charge
• Usage charge total
• Peak kWh
• Off-peak kWh
• Shoulder kWh
• Demand kVA
• Controlled-load kWh
• Network charges (where itemised separately)
• GST
• Total inc GST

A pivot on site code with billing period as the column dimension produces a cross-site, cross-period view directly. A pivot on vendor type compares electricity, gas, and water usage at the portfolio level. A pivot on retailer surfaces concentration risk, retailer-specific rate drift, and renegotiation candidates. The shape supports each of those without further reshaping.

Tenant recharge column shape. Tenant utility cost allocation Australia covers a spectrum — from full pass-through, where the entire bill flows to one tenant, to sub-metered arrangements where only the tenant's metered share applies, to percentage-allocated layouts where common-area costs are split by lettable area. The recharge view carries pass-through components broken out at the level the recharge agreement specifies:

• Site code
• Supply address
• NMI / MIRN / property reference
• Tenant code (recharge target)
• Recharge basis (full pass-through, sub-metered, percentage allocation)
• Billing period start
• Billing period end
• Days in period
• Supply charge (pass-through portion)
• Peak kWh and rate (where pass-through pricing applies)
• Off-peak kWh and rate
• Shoulder kWh and rate
• Demand kVA and rate
• Controlled-load kWh and rate
• GST on recharge components
• Total recharge amount

Where the recharge basis is full pass-through, the supply charge and usage charge components flow directly from the analytical row. Where it is sub-metered, the per-tenant kWh comes from the sub-meter reading and the rate columns supply the pricing applied. Where it is percentage-allocated, the unified row's supply and usage components multiply through an allocation percentage to land per-tenant amounts. Holding the rate columns alongside the kWh columns means the recharge calculation runs from the row itself rather than requiring a separate rates table.

The two variants are not exclusive — a portfolio that does both cost reporting and tenant recharge maintains the analytical row as the master, and the recharge view is a derived projection of it. Address normalisation belongs in either layout: the supply address column carries the canonical form (decided once for the portfolio — for example, unit-then-street with no abbreviations), and the join key column carries NMI, MIRN, or property reference so cross-period and cross-retailer joins remain reliable when the retailer's address string drifts.

The prompt names every column on the list, the AI extracts each field from whichever vendor's PDF arrives, and the output is the unified row in either of the two shapes. The shape is the contract; the prompt is how it is enforced.

---

From Extraction to Outputs: Cost Reporting, Tenant Recharge, and ESG Disclosure

The unified per-site row exists to feed downstream jobs, and the value of holding the row in the shape just specified is that those jobs become near-mechanical rather than a quarterly reshaping exercise.

Cost-per-site and cost-per-period reporting. The analytical row is already in the shape a portfolio review or board report wants. A pivot on site code with billing period as the column dimension produces the cross-site, cross-period view directly. Year-on-year comparison runs as a second pivot — same period last year against current period, with variance flagged where the difference exceeds a threshold. Procurement consultants use the same data to identify sites where tariff renegotiation would pay off: sites where demand kVA has crept up over the year and a demand-tariff renegotiation might be due, sites where peak / off-peak / shoulder usage has shifted such that a different time-of-use plan would be cheaper, sites where supply charge has drifted relative to usage and a flat-rate plan might fit better.

Tenant recharge. For commercial property and franchise networks, the recharge view feeds the tenant invoicing workflow each cycle. Where the recharge basis is full pass-through, the supply charge and usage charge components flow directly from the unified row to the tenant invoice. Where the basis is sub-metered, the per-tenant kWh from the sub-meter combines with the rate columns on the unified row to produce the tenant's amount. Where the basis is percentage-allocated, the unified row's per-site components multiply through the allocation percentage. The recharge calculation runs from the row, the GST treatment carries through correctly because GST sat as its own column, and the tenant invoice ties back to the source bill via the join key column.

Scope 2 emissions reporting. The kWh data on the unified row — peak, off-peak, and shoulder kWh combined into total kWh per site per period — is the input for Scope 2 emissions calculation under the GHG Protocol's location-based and market-based methods. Climate Active and the Clean Energy Regulator publish the emissions factors used to convert kWh to tCO2-e, and the National Greenhouse Accounts Factors document is updated annually. SMBs facing supply-chain disclosure pressure from larger customers — the kind that increasingly require Scope 2 numbers from suppliers as a condition of contract — find the per-site, per-period kWh column is already the right shape for the Scope 2 input. The methodology choice (location-based versus market-based, which factor set to apply, whether to include Scope 2 alongside Scope 1) is a separate subject; the data shape this article produces is the input.

The pattern of extracting documents into a per-site-and-period row that feeds reporting, recharge, and tax-claim downstream jobs runs across other AU document types too. Extracting Australian fuel invoices for fuel tax credit claims is the parallel where the source documents are fuel invoices and the destination row shape is built around the FTC claim — different document type, different destination, same workflow shape with the prompt as the contract between input and output.

Where vendor portals already produce structured exports — NEM 12 from AGL, account-summary CSVs from various retailers, online-portal usage histories where they exist — the answer is to use those exports and skip the extraction step. The extraction workflow is the path when those exports do not suit the job at hand: multi-vendor portfolios where retailers differ across sites and there is no single export that covers them all, bill-level data rather than meter-data-level (which excludes NEM 12 and similar usage exports), and the only-PDF-by-email reality for the many bookkeepers and property managers servicing portfolio clients without portal access. The workflow answers that reality. It is not a replacement for vendor exports where they exist and fit.