The intelligence platform for supervisory boards, executives and their insurers
Wirecard. Grenke. Adler. Signa. Every one of these crises had lead time — but nobody added up the weak signals. Foresain adds them up.
Reads the sources, extracts facts — each with URL and verbatim quote. A second, independent step verifies.
Versioned formulas compute score, rating and red flags — deterministic code, no AI involved.
WORM snapshots freeze every source with a checksum. Nothing is overwritten, everything stays traceable.
- Know before the AGM what the activist investor will ask - Preparedness instead of prediction - Board-ready reports
Foresain for executives →- Duty of care built on evidence, not slides - Red-flag screening with full source chains - Defensible on a Friday night
Foresain for supervisory boards →- A second data point next to the audit opinion - Continuous & standardised - Actuarially compatible
Foresain for insurers →- The layer between audit opinion and management report - Neither snapshot nor opinion - Measured, documented reality
Foresain for auditors & IR →Nine external modules are available today — calibrated against hand-researched reference cases. The remaining modules of the standards map are in preparation.
Composition, board evaluation, executive turnover, discharge and say-on-pay votes — the core of the standard.
Shareholder structure, AGM processes, voting-rights conduct, investor dialogue.
ICS/RMS architecture, three-lines model, effectiveness of control functions.
Related-party transactions, officer dealings, handling of conflicts of interest.
Sanction and proceedings history, CMS validation, management supervisory duties.
Audit committee, auditor rotation, enforcement and restatement history.
Remuneration system, target architecture, say-on-pay history and transparency.
Early risk detection and aggregation, liquidity provisioning, going-concern risk.
Guidance quality, ad-hoc discipline, ratings, short-selling exposure — the quantitative outside view.
Double materiality, supply chain due diligence, climate targets, greenwashing risk.
Accountability, approval processes and transparency in corporate AI deployment.
Incident history, board involvement, certification status — and the personal liability of officers.
Inside-view assessments based on internal documents and interviews — mandate only.
Disclosed formulas, rules and versioning — the auditable foundation of every module.
Same sources, same result. Every score is computed from versioned weights, anchors and rules — agentic at the edge, deterministic at the core. Two runs over the same evidence snapshot yield identical results, to the decimal. Every number in the report can be recalculated.
This rests on a strict division of labour: language models research and extract — scoring happens exclusively in reviewed, versioned code. Thresholds (such as the 80-percent discharge rule), red-flag criteria and weights exist as rules, and every module release carries a version number. A pinned-corpus mode replays any historical run exactly — the basis of every calibration and every audit.
No claim without a source. Every indicator carries its own evidence register: reference, original quote, evidence grade. The Evidence Confidence score openly states how robust the data basis is — and reports are stored as self-contained WORM documents: immutable, audit-proof, archivable.
Every source is preserved before processing as a snapshot with a SHA-256 checksum. A second, adversarial verification agent then tries to refute each claim against that snapshot — it can only discard evidence, never add it. Whatever fails verification drops out of the evidence base; missing information is reported as a gap, never estimated.
One GDS from 0–100, four rating bands, clearly named red flags. Supervisory board, D&O insurer and auditor read the same document — no translation layer. Assessment relies exclusively on public sources as of the reporting date, with a disclosed methodology aligned with DIN SPEC 33456.
The bands are fixed: A – robust (below 25), B – monitor (below 45), C – elevated (below 65), D – acute (65 and above); each red flag adds five points to the GDS. A red flag is a defined event — such as a discharge vote below 80 percent — not an editorial judgement. Every report carries an abstract, a management summary and an indicator index: from overall verdict to original quote in two steps.
Hosting and processing in Germany (Hetzner, ISO 27001 certified data centers), no cookies, no tracking, no dependency on US hyperscalers. At Foresain, data protection is an architecture decision, not an appendix.
Production runs in the Falkenstein/Vogtland data center under an Art. 28 GDPR data processing agreement; fonts and assets are served from our own infrastructure. Sovereignty extends to the assessment itself: Foresain evaluates public sources only — clients upload no internal documents, so no data-room and no confidentiality risk arises.
Governance risk does not emerge on the balance-sheet date. Foresain measures continuously against the same versioned yardstick — every change becomes visible as a delta before it reaches the annual report. The yearly exercise becomes an early-warning system.
In the portal every measurement becomes part of a time series: GDS trend per module, delta against the previous run, freshness labelling of the source base, peer comparison within the sector. Because the yardstick is versioned, every delta reflects a change in the company — not in the method. The median lead time measured across case studies: 14 months ahead of the visible event.
The fastest way to understand Foresain is a pilot: one company, one module of your choice, one complete evidence-grade report. Tell us which scenario you want to examine — we will get back to you within two business days.
hello@foresain.com
Foresain is a product of CRAiD GmbH. Assessments rely exclusively on public sources; hosting and processing in Germany.
Bel Foresa Vorherse n // Belegen n // Foresain / in // Vorhersehen rsehen // Belegen // elegen // Foresain // Vo oresain // Vorhersehen // B Vorhersehen // Belegen // Fore // Belegen // Foresain // Vorher // Foresain // Vorhersehen // Beleg n // Vorhersehen // Belegen // Foresain sehen // Belegen // Foresain // Vorhersehe legen // Foresain // Vorhersehen // Belegen / resain // Vorhersehen // Belegen // Foresain // orhersehen // Belegen // Foresain // Vorhersehen // // Belegen // Foresain // Vorhersehen // Belegen // Fo // Foresain // Vorhersehen // Belegen // Foresain // Vorh // Vorhersehen // Belegen // Foresain // Vorhersehen // Bel ehen // Belegen // Foresain // Vorhersehen // Belegen // Foresa egen // Foresain // Vorhersehen // Belegen // Foresain // Vorherse esain // Vorhersehen // Belegen // Foresain // Vorhersehen // Belegen rhersehen // Belegen // Foresain // Vorhersehen // Belegen // Foresain /