IPAC reads operational, financial, and external signals continuously, and returns the one thing every other system leaves to you: what changed, what it puts at risk, and what to do about it while there is still room to act.
Powered by Schubring Global Solutions
Three conditions require attention.
Current operations are stable; forward risk is elevated. Supplier certification failure, concentration exposure at a single processor, and shipments held at customs are driving the posture.
Resolve supplier certification failure
3 suppliers lack valid certification evidence on file, creating a material compliance and continuity exposure.
Temporarily restrict new orders from uncertified suppliers while updated evidence is collected.
Potential compliance failure, shipment hold, and customer assurance risk.
What changed · since yesterday
What IPAC is watching next
IPAC Command Center. Representative interface, demonstration data.
Nobody in a consequential operating environment is short of data. The regulatory feed arrives. The incident report arrives. The counterparty filing, the weather model, the supplier document. All of it arrives, most of it on time.
What does not arrive is the judgment. Which of those things actually changed something. What it puts at risk. Whether it can wait until Thursday. And when the call is questioned six months later, what the reasoning was at the time.
A dashboard hands that work back to the operator, arranged more neatly. It moves the problem; it does not finish it. So organizations buy another one, and the number of screens goes up while the number of defensible decisions stays exactly where it was.
More signal has never once produced more clarity. IPAC is built the other way round: it starts from the decision that has to be made and works backwards to the evidence that justifies it.
The output is the decision, not the dashboard.
Every domain on IPAC runs the same path. It is why a call in a supply chain is comparable to a call in a portfolio, and why the platform has not turned into six unrelated products behind one login.
Observations arrive continuously and fold into one evolving thread per condition. Not a new alert every day for the same problem.
government · commercial · your own record
Each thread is bound to the entities it actually touches: this supplier, this asset, this mandate, this counterparty.
entity graph
Scoring frameworks convert the condition into what it puts at risk, against a baseline for that domain.
scored against baseline
A specific, time-bound action, with the confidence behind it and the cost of not taking it.
confidence · risk of inaction
The action, its inputs, and the records it relied on are captured as one versioned decision record.
versioned · reviewable · durable
The gap is rarely that a signal was missing. It is that the signals were individually unremarkable, arrived on different days, and nobody was holding them together.
Signal correlations
Supply Chain Disruption Risk
CRITICALWeather event affecting facility operations combined with transportation delays.
Activate contingency suppliers and notify downstream customers.
Margin Pressure Risk
HIGHCommodity price movement combined with logistics disruption may compress margins.
Review contract pricing terms and hedge exposure.
What IPAC is watching next
Neither correlation above would have raised an alert on its own.
Predictions surface. Representative interface, demonstration data.
These are scored heuristics over your own graph, not a forecasting model with a track record. The recovery windows are industry ranges and are labeled as such in the product. IPAC will tell you what is converging and how much room you have left; it will not tell you what Tuesday holds.
What comes out
Not a chart that has moved. A named action, on named entities, with the confidence behind it, the cost of waiting, and the records it stands on. Four domains, one decision architecture.
Three lots have sat at customs past the third day, one of them short a certificate of analysis. IPAC names the lot the broker is actually waiting on.
Clear shipments held at customs
DEC-1191 · Guayas corridor · broker Andes Logistics
3 shipments have been held for three days or more, accruing delay and demurrage exposure. One lot is missing a certificate of analysis, which is what the broker is waiting on.
Escalate documentation to the broker and confirm release requirements per lot, starting with LB-2291-C.
Continued hold accrues storage cost and threatens downstream delivery commitments.
Held lots
84%· Moderate| Lot | Held | Docs |
|---|---|---|
| LB-2291-CPlanta Vinces 04 | 6d | 3 / 4 |
| LB-2288-APlanta Vinces 04 | 4d | 4 / 4 |
| LB-2274-FCoop. Guayas | 3d | 4 / 4 |
Every lot carries a public verification address, so a buyer can confirm origin and certification without an account.
A storm track has shifted east on three consecutive advisories. The decision window closes at 18:00 UTC, and IPAC says so before it does.
Hurricane Bianca — Gulf Coast
Seventeen assets fall inside the 72-hour cone, and two operators have already begun evacuation. The forecast has shifted east on each of the last three advisories.
Reduce Sabine Pass exposure and reroute Henry Hub nominations before 18:00 UTC.
A shut-in ordered after the window closes forfeits the reroute and leaves the position exposed through landfall.
Signal thread
87%· Moderate· ▲ worseningGulf Coast tropical system — Meridian Gulf assets
first Aug 17 · latest 12m ago · 9 observations
Nine observations across four sources, folded into one evolving thread. Not nine separate alerts about the same storm.
Two exposures that are survivable alone now compound inside the same window. The board sees them as one briefing, with the analysis attached.
Board briefing — Cascade County Employees’ Retirement System
Situation briefing for 12 September 2026
Posture Guarded · 22 days to meeting · est. funded ratio 78.4%
Funded ratio estimated from the last actuarial valuation rolled forward on reported returns. Not an actuarial opinion.
The plan is not in distress, but two exposures now compound: a fixed income mandate that has trailed for four quarters, and a January window where private capital calls land on the heaviest benefit outflow of the year. Neither is urgent alone. Together they narrow the board’s options if either moves before the next meeting.
Top three risks for the board’s attention
Core fixed income mandate has trailed benchmark for four consecutive quarters
Cumulative shortfall of 310bps against the Bloomberg Aggregate since Q3. Tracking error is inside guidelines, so this is a selection outcome rather than a mandate breach.
Δ since last meeting: +90bps shortfall since last meeting
Private markets capital calls collide with the January benefit payment cycle
Three funds have signaled calls inside the same 30-day window that carries the largest benefit outflow of the year. Unlevered cash covers it; the rebalancing sleeve does not.
Δ since last meeting: New since last meeting
Custodian and two managers share a single reporting vendor
A vendor outage would degrade valuation reporting across 38% of plan assets at once. No incident has occurred. The concentration itself is the exposure.
Δ since last meeting: Unchanged since last meeting
Decisions for this meeting
Core fixed income mandate — continue, watch, or put on notice
Recommend: place on formal watch with a two-quarter review trigger rather than terminate into a widening spread environment.
14 supporting records · analysis attached
A control change filed in another jurisdiction lands inside a line already at 94% utilization. The exposure is read across the book, not one name at a time.
Halvorsen Trading
Active · onboarded 2024-03-11 · reviewed 2026-05-02
A control change filed in Malta moves 22% of the ownership chain behind an entity that shares a director with a counterparty already on the restricted list. Utilization reached 94% of limit in the same week.
Freeze new lines at the current notional and require refreshed beneficial-ownership evidence before the next roll.
An unreviewed control change inside a 94%-utilized line is the shape of exposure that is discovered after it becomes a loss, not before.
Exposure by counterparty
78%· ModerateLimit utilization against the approved line. Concentration is read across the whole book, not one counterparty at a time.
Representative IPAC interfaces reproduced from the shipped product. Demonstration data throughout.
The engine is shared: ingestion, entity context, scoring, decisioning, evidence. What differs by domain is the baseline it scores against and the shape of the decision at the end.
Also running on IPAC: Family Office Intelligence · Reinsurance Intelligence
The architecture in detailBuilt for scrutiny
Anything IPAC recommends can be opened. The factors that produced it and their weight. The sources, and when each was last retrieved. The confidence, stated as a number. The documents cited, with their digest, their version, and their review state, including the one that is missing.
The record is versioned, so superseding a document does not rewrite the decision that relied on the earlier one. Six months later the question “what did we know, and when” has an answer that does not depend on anyone’s memory.
Basis · Resolve supplier certification failure
DEC-1184 · 2026-08-20 06:14 UTCAppend-only SHA-256 chain. Each entry chains to the previous, so a later alteration is detectable. Internal, not a public ledger.
Decision basis view. Representative interface, demonstration data.
A buyer, an auditor, a regulator. IPAC publishes a single claim at an unguessable address: origin, lot, plant, certifications, document digest. Nothing else about the tenant it came from. Verification tokens are never returned by list endpoints without authorization.
No account. No tenant data. One claim.
Public verification page at /verify/:token.
Independently examined under SOC 2 Type 1 and Type 2 by Sensiba.Type 2 period March 5 to June 5, 2024Scope and limits
IPAC is developed by Schubring Global Solutions, whose work sits across security, risk, intelligence, and technology for organizations where being approximately right is not the same as being defensible.
That shows up in the product as a bias: state the limit instead of softening it, attach the evidence instead of asserting the conclusion, and never claim a capability the codebase does not have.
About IPAC and SGSWe will walk it end to end: the signals that would have surfaced it, the exposure it carries, and the record the call would have left behind. If IPAC would not have caught it earlier, we will say so.