Asil Operational Intelligence READ · PLATFORM ENERGY-NL·ES / 5-STACK · JUL 2026

Prepared for an operating-principal lens · Benelux

We read your newest platform before you sent us anything.

Everything below in blue is public record — announcement dates, entity structure, the law and its deadline. Everything in amber is illustrative until one sanitized export replaces it. Nothing here required access, credentials, or a single non-public byte.

public record — verifiable illustrative — until your export every € computed, never estimated

Cyberbeveiligingswet · in force 15 Aug 2026

— d — h — m — s

Passed the Eerste Kamer 7 July 2026. Energy is one of the 18 sectors in scope. public record

Fines to €10M or 2% of global turnover · personal board liability · 24h/72h incident clocks

The platform, from the public record only

19 JUN 2025 public record

Four Dutch energy-software companies — forecasting & balancing (est. 1997), retail & billing, flexibility & trading, grid data — fuse under a single new name.

30 MAR 2026 public record

An Iberian add-on joins from Tarragona — market communication, contracting, forecasting, billing for 100+ Spanish suppliers. The announcement itself says integration “will take place over the coming months.”

07 JUL 2026 public record

The Eerste Kamer passes the Cyberbeveiligingswet — the Dutch NIS2 implementation. Registration duty, duty of care, incident reporting.

15 AUG 2026 public record

In force. The platform's customers are essential entities; chain responsibility reaches their software suppliers. Group-level evidence, not five separate stories.

Five stacks. Two jurisdictions. One month. We never named the platform — you recognized it. That is the entire product: reading what the seams already say.

The seam map

structure from public record · signals illustrative

revenue collision duplicate spend compliance seam

Nothing migrates. The stacks stay exactly where they are — the read runs across them, as-is, from one sanitized export per entity.

Now run the read yourself.

The fragments on the left are sanitized stand-ins for what five stacks emit every week. Edit any number — the exposure recomputes live, because the engine owns every figure; the model only reads. On a call, we clear these cards and paste yours.

Raw signal — editable · five stacks

Unstructured fragments, exactly as they arrive. Numbers are highlighted because they are the only thing the engine will compute with.

Try it: change €140k to €300k in the first card, run again — every downstream figure follows.

🔒 People data enters as anonymized aggregate only. This file makes no network calls and contains no credentials — open the source and check.

Intelligence

Observed → inferred → prepared action. Confidence is a stated function of corroboration — not a vibe.

◦Mapping five entity schemas — nothing migrates

◦Resolving shared accounts across stacks (VAT-root · domain · invoice refs)

◦Sweeping vendor ledgers for cross-stack overlap

◦Checking group evidence coverage against the 15 Aug duty set

Five dashboards, all green — until you run the read across them.

—

Cross-stack findings

—

Risk

—

Exposed

forwardable one-pager — all figures computed, assumptions labeled

📤 0 action(s) queued — a person approves before anything runs

▸ Assumption table — the only numbers we bring

€/hr loaded (NL mid-market) = 90 · weeks/month = 4.3 · working days/yr = 260
Everything else is read from your fragments. Change a rate here on a call; every figure follows.

Recommended sequence — next 14 days

Don't believe this file. Falsify it.

Two queries your platform CFO can run internally, this week, without us. They cost a day and settle the question.

Query 1 · revenue seams

Top 20 accounts appearing in ≥2 of the five stacks → cross-check against (a) open support escalations and (b) active expansion quotes. Count the collisions.

Query 2 · spend seams

One consolidated vendor list across all five entities → flag vendors appearing ≥2× with no group agreement on file. Sum the annual delta.

If both come back clean — delete this file, and our apologies for the ten minutes. If either can't be produced in a day, that inability is the finding.

GATE 1What “one sanitized export” actually means

Per entity: account list (name, VAT, domain), open tickets (status + age), active quotes (value + stage), vendor ledger (name + annual). CSV or native export, ≤ 2 hours of one admin's time per entity. No API keys, no live credentials, no agents inside your systems — ever.

GATE 2Entity resolution, shown not asserted

Matches score on VAT-root + domain + invoice references. ≥ 0.90 auto-links; below that, a human reviews — you saw a live match card above. Precision is reported per run; a wrong link can pause real revenue, so the threshold errs conservative.

GATE 3Data terms, before data

DPA signed before a single row moves. EU processing only. Read-only by construction. The model reads structure and text; the deterministic engine owns every number. Deletion on request, evidenced. This demo file itself ships with zero credentials and zero network calls.

Six weeks. One export. Real numbers before Q4 planning.

A paid summer pilot on your messiest platform: we replace every amber figure in this file with your computed ones, or we tell you the seams are clean — both answers are worth having before the board asks.

Scope

One platform · five stacks · read-only

Input

One sanitized export per entity

Duration

6 weeks · fixed fee

Success

≥3 verified findings, or the NIS2 evidence map

Governance

DPA first · EU-only · person approves every action

Timing

Results land before Q4 planning