Product · four modules

Ingest. Extract. Review. Export.

No ERP connectors. No 6-month implementation. Four document-driven modules that turn the documents you already have into reconciled, audit-ready data. The report on top is a template, so the same four modules drive every output the engine produces.

01 · Ingest

Upload what you already have.

PDFs of energy bills and supplier invoices. Scanned and digital PDFs both work. Last year's filing as a reference. Drag a folder, get a queue.

  • PDF · digital and scanned (image-based), with OCR built in (runs locally)
  • Excel (.xlsx), with every sheet treated as a page
  • Word (.docx)
  • CSV with auto-delimiter detection
  • Swiss-apostrophe thousand-separator handling (CHF 4'610.00)
  • One workspace per filing entity: a single company or one client
upload queue · acme-202414 files
energy-bill-q4.pdfparsedgrounded
payroll-summary-2024.pdfparsedgrounded
supplier-emissions.pdfparsedgrounded
prior-year-report.pdfqueued
travel-summary-2024.pdfneeds reviewverify
02 · Extract and ground

Every figure pinned to the line it came from.

The local LLM reads the figures out of your source documents and pins each one to the exact page it came from, then the verifier cross-checks it against that page before it ships. When the verifier can't ground a figure it refuses rather than inventing one, and routes it to a reviewer with a verify chip. That refusal is the point: it is what makes the output safe to put in front of a client or an auditor. For a fiduciary this produces the reconciled Operating Report. On the named-account path, where the target is ESRS, the full EFRAG IG 3 set is catalogued (1,144 datapoints) and 48 indicators are auto-extracted with citations today, with narrative drafted at the section level. Per-datapoint narrative is not shipped.

  • Every figure cited to its source page, in any supported document
  • 48 ESRS indicators auto-extracted with page-level citations named-account path
  • 1,144 EFRAG IG 3 datapoints catalogued named-account path
  • Section-fan-out: one focused prompt per ESRS section, run in parallel for accuracy
  • Coverage tab with filter chips (gaps · mandatory gaps · approved · covered)
  • Per-datapoint narrative drafter on the roadmap roadmap
  • Re-runs are idempotent; old citations preserved in the audit log
ESRS E1 · para 3page-level verifier · v2024.11
E1-5 · energy1,200 MWhgrounded
E1-6.1 · Scope 1412 tgrounded
E1-6.2 · Scope 21,108 tgrounded
E1-6.3 · Scope 314,000 tverify
E1-7 · removals0grounded
03 · Review and approve

Assign. Approve. Sign off.

Multi-reviewer flow: assign reviewers per file, approve per metric, sign off per report section. External auditors get a read-only role and a persistent banner telling them so. Every action is logged with named approver and UTC timestamp in an append-only audit log.

  • Per-file reviewer assignments
  • Per-metric approval with named approver; only approved figures reach the export
  • Per-section sign-off, carried into the export
  • External auditor · read-only role enforced on every review, approve and export action
  • AI suggestion popovers (Apply / Dismiss) for ungrounded values, code re-mappings, and narrative edits
  • Append-only audit log with a per-entry sha256 hash chain
sign-off · ESRS sections2 / 12 complete
E1 · climate changesigned · 17:42 UTC
E2 · pollutionsigned · 17:48 UTC
E3 · waterpending review
S1 · own workforcepending review
G1 · business conductdraft
04 · Export

Six formats. One bundle that auditors can verify offline.

The evidence pack ships as a zip with a manifest.json carrying SHA-256 per source file and a chain_sha256 per metric, so auditors re-verify the bundle months later without needing the platform. The CSRD pack also bundles an append-only audit_log.json with its own hash chain.

  • A4 PDF with optional branded logo
  • iXBRL XHTML preview filing-grade taxonomy on roadmap
  • ESEF report-package zip
  • Evidence-pack zip · SHA-256 manifest + per-metric hash chain
  • CSRD JSON for downstream tooling
  • Annual Income & Reconciliation XLSX accountant pipeline
evidence-pack.zip · acme-2024SHA-256 manifest
manifest.jsonverified
audit_log.jsonchain ok
report.pdfA4 · 84 pages
report.xhtmliXBRL demo
sources/ (14 files)all hashed
What's deliberately not here

No ERP connector. That's the point.

Attera does not plug into SAP or Snowflake, and it does not need to. It runs on the documents you already produce: energy bills, payroll exports, supplier invoices, last year's report. You upload them, and that is the whole integration.

See it on your own documents.

Bring a sample of your real source documents to the call. We'll generate a draft live and you'll see the citations land.

Book a demo