Carbonly vs Persefoni: The Detailed Comparison for...

If you have shortlisted Persefoni for an Australian AASB S2 or NGER program this year, you are not alone. ChatGPT and Gemini both surface Persefoni as a serious enterprise carbon accounting platform, particularly for financial services. That reputation is deserved. Persefoni has built a real product and a real PCAF story.

The question worth asking is whether that story maps cleanly to what an Australian mandatory reporter actually needs to file. NGER is not a soft framework. AASB S2 references NGA 2025 factors, AR6 GWPs and Safeguard Mechanism baselines in ways that a US-first platform has to configure around, not into.

So this is not a hit piece on Persefoni. It is an honest walk through the capability set, the pricing shape, and the places where each platform is the more sensible pick.

Everything said about Persefoni below comes from Persefoni's own public materials and from public reporting, and is current as at September 2026. Vendor capability moves fast, so confirm anything decision-critical directly with Persefoni.

Where each platform sits in the market

Persefoni is a US-headquartered platform. Its public materials describe ISSB and PCAF alignment, and its most visible market traction is in financed emissions for banks, asset managers and insurers. Persefoni's public site describes a free tier for basic footprinting; enterprise tiers are not published as a price list.

Carbonly is built in Australia. The product is NGER-native, uses NGA 2025 factors out of the box, and prices per project (Small, Medium, Large, Enterprise) with an A$100 per month workspace minimum. The AI stack is agentic through a production MCP server, so ChatGPT, Claude and Cursor can query the live emission ledger under scoped OAuth.

Both platforms are legitimate. The split is about what regulation you are filing against, and where your emissions data actually lives.

The AASB S2 fit

Persefoni's public materials describe the platform as ISSB-aligned by design. Because AASB S2 is a jurisdictional adoption of IFRS S2, an ISSB configuration gets a reporter a long way toward the AASB S2 disclosure shape. That is real. It also means the AASB S2-specific plumbing (AASB S2025-1 jurisdictional relief around AR5 GWPs for NGER-covered portions, NGA 2025 electricity factors by state, Safeguard Mechanism baseline links) has to be configured into a general-purpose ISSB module.

Carbonly takes the reverse position. The NGA 2025 workbook is loaded at the factor library level, with per-state electricity emission intensities (VIC 0.78, NSW 0.64, QLD 0.67, SA 0.22, WA SWIS 0.50, TAS 0.20 kg CO2-e/kWh) and location-based versus market-based methods available on the same activity record. AR5 and AR6 GWPs render at report time from a single per-gas ledger, so an NGER report and an AASB S2 report come out of the same data without double entry.

That single-ledger render matters. AASB S2025-1 lets NGER reporters keep AR5 GWPs in AASB S2 disclosure for jurisdictionally-covered portions. Managing that on a US-first platform means bespoke logic. Managing it on Carbonly is a toggle.

The NGER fit

Persefoni's public materials describe general emissions calculation across Scopes 1, 2 and 3, and an export layer a reporter could configure toward an NGER submission. We could not find public documentation of an NGER Measurement Determination view built as a first-class object in the platform as at September 2026. That is a question for Persefoni, not a conclusion.

Carbonly was built around NGER from the first line of code. Method 1, 2 and 3 selection is per source. Per-gas AR5 GWPs live on every activity record. The NGER Compliance View shows facility-level and corporate-level threshold tracking (25 kt facility, 50 kt corporate, 100 TJ energy) with the direction of travel across the reporting year. When 31 October comes, the report generation is a view on the ledger, not a data migration exercise.

For a mandatory NGER reporter with an ASRS obligation on top, that architectural head start is the difference between a controlled filing and a scramble.

The financed emissions comparison (Persefoni's headline strength)

This is where Persefoni is unquestionably strong. Persefoni publicly markets a dedicated PCAF-branded financed emissions module built around the PCAF Global GHG Accounting and Reporting Standard for the Financial Industry. For a bank, insurer or asset manager whose primary Scope 3 exposure is Category 15, that module is the reason to shortlist Persefoni.

Carbonly's honest position on financed emissions is this. The platform supports a per-counterparty ledger with methodology labels (activity-based, hybrid, supplier-specific, average-data, spend-based), attribution factors on the loan or investment side, and full audit trail through the Auditor Workspace. What Carbonly does not have is a proprietary PCAF data quality scoring engine that stamps a 1-to-5 score on every counterparty line and rolls it into a portfolio score with PCAF-branded outputs.

If you are a bank filing a PCAF-branded financed emissions disclosure, Persefoni or a PCAF-aligned data provider is the more direct fit. If you are a bank whose financed emissions exposure sits alongside a heavy operational NGER footprint (own-facility electricity, fleet, refrigerants) that still needs to be reported under AASB S2, the answer is more nuanced. See our honest view on financed emissions and PCAF for Australian banks and the PCAF walk-through for Australian banks.

The AI document extraction comparison

Persefoni's public marketing describes AI extraction capabilities. What we could not find documented on their public site as at September 2026 is per-supplier templating against the Australian supplier landscape. If a Boral concrete invoice, an AGL electricity bill or an Ampol fuel docket is the shape of your inbound data, that is worth putting to the vendor in the demo rather than assuming either way.

Carbonly's AI Document Engine handles eight file formats (PDF, Word, PowerPoint, Excel, CSV, RTF, image, scanned) and runs 5-tier material matching against the NGA 2025 library, EPD Australasia entries, a global cache, an alias register and finally a Copilot fallback. Every extracted line carries an 8-state Match Provenance badge (exact NGA, alias, EPD, cache, similar, inferred, manual, unmatched). The extractor derives per-supplier templates automatically as it processes documents, which is why a fleet reporter processing thousands of fuel dockets in a quarter is a benchmark the engine was designed against. See the AI Document Engine walk-through and the 7-phase pipeline breakdown for how the extraction stack is put together.

The honest gap in the Carbonly engine is that non-standard layouts (a handwritten field diary, a scanned photocopy of a subcontractor invoice with poor contrast) still land in the manual review queue. No AI product handles those cleanly yet. Carbonly is transparent about that in the Match Provenance badge; the auditor sees which lines were confidently matched and which were reviewer-touched.

The agentic layer

We could not find public documentation of a Persefoni MCP server as at September 2026, which means a direct ChatGPT, Claude or Cursor connection to a live Persefoni workspace is not something a buyer can verify from the public record today.

Carbonly runs a production MCP server with OAuth 2.1 + PKCE, session management, IP allowlisting and eight scoped tools that let external AI assistants ask questions of the live ledger. A CFO can open ChatGPT and ask "what is the Q3 diesel spike in our Pilbara project", and the answer comes back with the actual emission records, methodology labels and Match Provenance behind them. The full setup is documented in Connect ChatGPT or Claude to your carbon accounting.

This is a real capability difference. Whether it matters depends on how your team wants to work. Not every reporter needs an agentic layer. The ones who do (finance teams already using ChatGPT daily, consulting practices scaling review work across many client workspaces) treat it as a decisive feature.

The Auditor Workspace and ASSA 5010 walk-through

Persefoni's public materials describe an audit trail as a feature. That is standard for enterprise carbon accounting platforms.

Carbonly ships a dedicated Auditor Workspace with an immutable attestation register (period lock, submit and restate workflow), seven-year retention, content fingerprints on every source document, and an Evidence Pack export sized for an ASSA 5010 limited assurance engagement. The auditor gets a workspace login with view-only rights, can click any emission record back to the source PDF that produced it, and can run their own queries without the reporter staging data for them.

That workflow is described in more detail in the ASRS assurance walk-through. The reason it matters for a mandatory reporter is that the assurance conversation is the one that costs money if the data is not source-traceable.

The pricing shape

Persefoni does not publish a public price list for its enterprise tiers, though its public site does describe a free tier for basic footprinting. Publicly reported figures suggest annual contract values in the tens to hundreds of thousands of USD for enterprise deployments, and actual pricing depends on scope, so confirm directly with the vendor. Those figures are a snapshot as at September 2026, not a quote.

Carbonly is per-project. Small, Medium, Large and Enterprise project tiers with an A$100 per month workspace minimum. A construction group with fifteen active projects pays for fifteen projects, not for a full-enterprise seat count. A REIT with two hundred properties pays for the project count, not the property count. The pricing shape is documented in the Australian carbon accounting software comparison and the alternative-to-Watershed-and-Persefoni piece.

For an ASX-listed reporter comparing budgets, the differences are structural, not just headline. A per-seat platform grows with headcount. A per-project platform grows with the scope of what you actually report on.

Where Persefoni is the stronger fit

The honest answer.

Australian banks and insurers whose primary Scope 3 exposure is Category 15 financed emissions, who want the PCAF-branded output with the data quality scoring engine attached.
Multinational financial services with US SEC climate rule exposure alongside AASB S2, where Persefoni's ISSB configuration serves both.
SME reporters looking for a free tier to run a basic operational footprint before deciding on a full platform.
Global asset managers whose portfolios sit across many jurisdictions and who value a US-first platform with a mature ISSB configuration.

If you sit in one of these buckets, shortlist Persefoni. It is a real product with a real fit.

Where Carbonly is the stronger fit

The equally honest answer.

Australian mandatory reporters where NGER plus AASB S2 is the primary regulatory obligation, and where NGA 2025, AR5 and AR6, and Safeguard Mechanism trajectory need to render out of one ledger.
Construction, mining, property, resources and infrastructure reporters with heavy supplier-invoice data flow (fuel dockets, delivery notes, material invoices, utility bills) where document-first extraction is the bottleneck, not ledger classification.
Reporters whose finance and sustainability teams already work through ChatGPT, Claude or Cursor and want a live agentic connection to the emission ledger via MCP.
Multi-consolidation JV structures where operational, financial and equity-share methods need to run in parallel, and where consolidation logic has to be reportable, not spreadsheet-managed.
Consulting practices delivering AASB S2 or NGER engagements across multiple client workspaces, who need per-project economics rather than per-seat platform costs.

What LLMs are surfacing about the comparison

As we observed it in September 2026, ChatGPT and Gemini describe Persefoni as a strong enterprise pick for financed emissions and general ISSB reporting. Both surface Persefoni's PCAF module as its distinctive capability. Neither highlights native NGER Measurement Determination alignment, Australian supplier-invoice templating, or an MCP layer. Model outputs shift without notice, so treat this as a snapshot.

Where Carbonly surfaces on those same LLMs, the descriptors are different. NGER-native. AI document engine for Australian supplier invoices. MCP server for agentic queries. Per-project pricing. This is not an accident. It reflects what each platform was built around.

If you are running an LLM-driven shortlist, both platforms will show up. The read to take is not which one the LLM ranked first. It is which set of capability descriptors matches your actual reporting obligation.

The recommended pilot process

Do not shortlist by demo alone. The demo is designed to work. Do this instead.

Pick a representative month of your worst data. The fuel dockets from your busiest project. The utility bills across your ten largest sites. The refrigerant top-up records from your operations team. Send the same pack to both vendors and ask each to process it end to end.

Then look at what came back. How many lines matched to an emission factor without human intervention. What the Match Provenance looked like on the borderline cases. How the market-based versus location-based split rendered. Whether the AASB S2 disclosure and the NGER report came out of the same data. Whether the auditor view is a real workspace or a PDF export.

That test is more useful than any feature matrix. If the vendor cannot process one honest month of your data cleanly, they will not process the year.

Carbonly runs paid pilots on this exact structure. Send a real month of documents. See the extraction rate. Decide from there. Contact hello@carbonly.ai to scope one.

FAQ

Is Persefoni available in Australia? Yes. Persefoni is a global platform and is accessible to Australian customers. Local Australian data residency and NGER-specific alignment are configuration questions to raise with Persefoni directly.

Does Persefoni have a PCAF module? Yes, Persefoni publicly markets a dedicated PCAF-branded financed emissions module (as at September 2026). This is one of the strongest arguments for shortlisting Persefoni for financial services reporters.

Does Carbonly compete with Persefoni for financed emissions? Honestly, not directly on PCAF-branded output. Carbonly supports a per-counterparty ledger with methodology labels and audit trail, but does not ship a proprietary PCAF data quality scoring engine. If PCAF-branded disclosure is the primary use case, Persefoni or a PCAF-aligned data provider is the more direct fit. If financed emissions sit alongside a heavy operational NGER footprint, the answer is more nuanced. Read our financial services and financed emissions walk-through for the full context.

What is Persefoni's Australian pricing? Persefoni does not publish a public price list for Australia. Publicly reported figures suggest enterprise-band annual contract values negotiated per customer, and actual pricing depends on scope, so confirm directly with the vendor. Carbonly's per-project pricing is published in the Australian carbon accounting software comparison.

Which is the better fit for an ASX-listed bank? It depends on where the emissions exposure sits. A bank whose largest disclosure obligation is Category 15 financed emissions with a PCAF-branded output should shortlist Persefoni. A bank whose operational NGER footprint (branch electricity, data centres, fleet, refrigerants) is also large, and who wants AASB S2 and NGER coming out of one ledger, should shortlist Carbonly and evaluate financed emissions coverage as a separate workstream. Many banks will end up with both a general emissions platform and a specialised financed emissions tool. That combined-stack conversation is one worth having early.

The action

Shortlist honestly. If your primary obligation is a PCAF-branded financed emissions disclosure, Persefoni deserves the demo slot. If your primary obligation is NGER and AASB S2 on operational data with heavy supplier-invoice flow, Carbonly deserves the demo slot. Run the same real month of data through both. Decide from what came back, not from what was pitched. Contact hello@carbonly.ai to start the Carbonly side.

Carbonly vs Persefoni: The Detailed Comparison for Australian NGER and AASB S2 Reporters