How Palau works
One twin. Every answer.
Instead of a form to fill, Palau builds a governed digital twin of your organisation — and puts your report, and everything after it, on top.
Feed the Vault
Drop in policies, reports, evidence — anything. AI categorises every file, detects its reference year, entity and validity, and keeps versions for the audit trail. Nothing is used unless you say so.
Mirror your structure
Group, entities, sites — including partial ownership and joint ventures. The twin knows who owns what, so consolidation and materiality follow your real organisation, not a template.
Report on any framework
AI drafts every datapoint from your own documents, scores compliance paragraph by paragraph, and routes the rest to the right colleagues. You review, approve, consolidate, export.
Reuse everything
The same twin answers customer tenders, EcoVadis, CDP, benchmarks you against peers and screens acquisitions. This is where compliance starts paying you back.
Reporting, properly
The hard parts of reporting, handled.
Not a text generator with a checklist. Six capabilities built for how a real sustainability statement actually gets made — and audited.
AI auto-population, with receipts
One click drafts a full reporting project from your own documents. Every sentence cites its source — down to the page — is flagged as AI-generated, and counts for nothing until a human approves it.
Gap analysis in minutes, not months
Upload last year's report and get a compliance score per datapoint, paragraph by paragraph against the standard — with drafted improvement points. It keeps checking while you write.
Consolidation that knows your group
Roll up narrative and numbers from entity to group — including proportional consolidation for partial ownership. No central ERP required. Cordeel consolidates 25 companies across six countries this way.
A double materiality auditors can interrogate
Import your existing DMA or build one from scratch — scoring, thresholds, IROs, per-entity materiality. Every step is documented and stress-tested by people. No black box.
Contributors who aren't experts — by design
Colleagues see only their datapoints, in plain language, in their language, with a deadline and an AI assistant beside every question. They answer; they don't decode regulation.
An audit trail auditors compliment
Who changed what, when, and why — version history with exact diffs, evidence locked to datapoints, omissions justified. Give your auditor access, or let them question the AI's logic directly.
60%
less time on questionnaires and tenders — a KPI we commit to in contracts
1.5h
to answer a customer tender estimated at 4–6 weeks of manual work
25
companies consolidated across six countries — without a central ERP
48h
to reflect regulatory changes in the platform when the EU moves
Each datapoint gets its own agent.
It knows the requirements for that exact datapoint, and comes with prompts that take the manual work off your hands. You keep the judgment.
E1 · required · tCO₂e
Out-of-the-box prompts
In FY24, Group Scope 1 emissions were 12,238 tCO₂e, down 6% on FY23. The reduction came mainly from the Ghent site's switch to renewable process heat. (Draft, yours to edit.)
Logged to the audit trail
Every action, on the record.
A full audit trail sits behind every number: who changed what, when, and why.
Datapoint agent drafted the E1-6 narrative
just nowPriya N. set Scope 1 to 12,238 tCO₂e
2 days agoLena K. approved Climate & energy
3 days agoTomas R. commented on energy intensity
4 days ago
Report on your terms
A pilot sets up your frameworks, your structure, and your datapoints.