AI Usage ESG Reporting
From personal awareness to organisational accountability.
The Challenge
You understand AI's per-query environmental impact (thanks to our free calculator). But your organisation needs something different: a defensible, audit-ready line item in your Scope 2 / Scope 3 emissions inventory that survives regulatory scrutiny.
Generic spend-based factors (EEIO) overestimate AI emissions by 10–40×. Your CFO and assurer deserve better.
The Solution: Three Service Tiers
All tiers use the same methodology as our free calculator, anchored to first-party operator data, but tailored for organisational reporting.
Your AI usage is split by deployment mode: self-hosted models (Scope 2) vs. purchased third-party AI services (Scope 3, Category 1). The methodology statement is the product.
Generic spend-based (EEIO) factors overestimate AI emissions by 10–40×. This annex is anchored to first-party operator energy disclosures, the same methodology as the free calculator.
Mapped to Australia's NGER scheme and the GHG Protocol, with confidence bands and a versioned methodology statement your assurer can review.
Tier 1: Concierge Annex (AU$990)
- Deployment-mode split (Scope 2 self-hosted vs. Scope 3 Cat 1 purchased AI services)
- Physically-derived energy, water and carbon factors with confidence bands
- NGER + GHG Protocol-aligned methodology statement
- Delivered as a PDF annex, typically within 5 business days
Best for: Organisations reporting this year who want a quick, defensible baseline.
Tier 2 & 3: Self-Serve Platforms (Coming Q4 2026)
Automated annual reporting with API integration and audit trails for teams and consultants.
Tier 1 is done for you. Tiers 2 and 3 put the same methodology in your hands: a self-serve platform that turns your organisation's AI usage into an audit-defensible emissions annex, refreshed every reporting period. These platforms are in build and expected to open in Q4 2026. Join the waitlist to help shape them and secure early access.
Starter
Automated annual reporting for small teams.
or AU$1,490 billed yearly (about two months free)
Best for: A single organisation or small team preparing one annual submission.
- One legal entity
- NGER and GHG Protocol Scope 2 and Scope 3 classification and export
- 12 months of reporting history, so last year's figures reproduce
- Auditor-pack PDF and XLSX exports
Business
Multi-period tracking, API integration and audit trails.
or AU$3,990 billed yearly (about two months free)
Best for: Mid-market organisations reporting across multiple entities and frameworks.
- Up to 5 legal entities in one account
- All framework mappings: GHG Protocol, CSRD / ESRS E1, GRI 302 / 303 / 305, NGER, CDP and IFRS S2
- CSV import and multi-period tracking
- Uncertainty and confidence bands on every figure
- API integration and an append-only audit trail
Practice
For sustainability consultants serving multiple clients.
or AU$7,990 billed yearly (about two months free)
Best for: Sustainability consultants and assurers reporting on behalf of multiple clients.
Everything in Business, plus:
- 10 client entities in one workspace
- Co-branded annexes issued under your practice
- Per-client audit trails and exports for assurer review
- One dashboard across every client's reporting period
Pricing is indicative, shown in Australian dollars, and subject to change before launch whilst we validate what teams and practices are willing to pay. Joining the waitlist creates no obligation and reserves no charge: there is nothing to buy yet.
Every self-serve tier is built on the same engine
- Physically-derived factors, never spend-based (EEIO) proxies
- Append-only methodology audit trail: the assurance premium
- Confidence bands shown on every figure, never hidden
- AI water reporting mapped to GRI 303
- Server-side entitlements, so your data stays scoped to your account
- Exports as auditor-pack PDF, XLSX or structured JSON
- Connectors to feed Watershed, Persefoni, Sweep and Normative
Isn't AI usage too small to bother reporting?
Materiality is not the point: defensibility is. Once AI appears anywhere in your operations, an assurer will ask how the figure was derived. A spend-based (EEIO) estimate overstates AI emissions by 10 to 40 times and cannot be defended line by line.
These tiers give you a physically-derived figure with visible confidence bands and an append-only methodology trail, so the number survives review whether it turns out large or small. A small, defensible line item beats a large, indefensible one.
Join the Q4 2026 waitlist
Why Physically-Derived Beats Spend-Based
| Spend-based (EEIO) | Physically-derived (Ours) |
|---|---|
| ~1–2 kg CO₂/$ spent | 0.367 kg CO₂/kWh actual |
| Used as proxy across all software | Anchored to first-party operator disclosures |
| Overestimates AI by ~20–40× | Order-of-magnitude guide |
Result: Your assurer can defend the numbers.
Questions?
See our Methodology page for sources and assumptions.
Or email us to discuss your organisation's reporting timeline.
Every figure is an estimate based on available disclosures and modelled assumptions, not a metered actual, nor an assurance opinion. See the methodology page for sources and limitations.