Platform
Integrations
All use cases
IndustriesPricingProductField Notes
Sign inGet started
Otiox Command · Predictive Engine

When a business is modelled as it truly is, its future becomes legible.

Get startedView platform
Otiox holds your operation as an ontology — suppliers, materials, machines, orders and customers, and every relationship between them. So a late purchase order is never merely late: it is a stalled machine next week and a broken promise the week after. The Predictive Engine follows those consequences forward, and shows you which are taking shape, how likely they are, and why.
app.otiox.comFig. 00
Risk Inbox
What is likely to go wrong next, why the engine thinks so, and what you can do about it.
Run engine
Open · 14Acted on · 6Came true · 3Averted · 5
RecordRiskLikely
MO-PRED-0001Late production82 %
SO-PRED-0004Promise at risk71 %
LOT-2026-118Lot expiring64 %
INV-2026-0213Payment late58 %
RM-STEEL-04Stockout47 %
01

Foresight, with its reasoning shown.

A prediction you cannot question is a prediction you cannot trust. Every risk in Otiox names the record it concerns, the chain of records that leads to it, and the action most likely to change its course.
Risk Inbox
What is likely to go wrong next, why the engine thinks so, and what you can do about it.
Run engine
Open · 14Acted on · 6Came true · 3Averted · 5
RecordRiskLikely
MO-PRED-0001Late production82 %
SO-PRED-0004Promise at risk71 %
LOT-2026-118Lot expiring64 %
INV-2026-0213Payment late58 %
RM-STEEL-04Stockout47 %
01

Risk Inbox

Every open risk in one place — late production, promises at risk, expiring reserved lots, stockouts, late payments — ordered by likelihood, each one a click from the record it concerns.

Attention, spent where it matters

Causal Map
Each chain is a hypothesis until an investigation supports or refutes it.
ChainsDependenciesConfirmed links
Supplier delay → late productionConfirmedLinked recordsOrder in timeStrong
PO-PRED-0001
+33 d late
Shenzhen Precision
MO-PRED-0001
+24 d late
SO-PRED-0001
Machine down → late productionHypothesisLinked recordsOrder in timeModerate
CNC-01
Spindle bearing seized
MO-PRED-0001
+24 d late
SO-PRED-0001
02

Explained, not merely scored

Each prediction carries its path: this purchase order is 33 days late, it feeds this manufacturing order, which builds the goods for this customer. The reasoning is visible, so it can be examined.

Confidence earned, not asserted

What-if scenario
Changes apply to a copy of the plan. Nothing real moves until you say so.
IfShenzhen Precisionships+14 daysthen…
MeasureBaselineScenario
On-time customer orders91 %78 %
Manufacturing orders late27
Cash low point€184k€121k
Material to expedite—3 POs
Affected: 2 suppliers · 7 MOs · 11 customer ordersSave scenario
03

Forecasts grounded in relationships

Demand, cover and cash are forecast from what is actually connected: open orders, reservations, supplier lead times as they truly behaved, machine availability.

Forecasts that move as the business moves

What-if scenario
Changes apply to a copy of the plan. Nothing real moves until you say so.
IfShenzhen Precisionships+14 daysthen…
MeasureBaselineScenario
On-time customer orders91 %78 %
Manufacturing orders late27
Cash low point€184k€121k
Material to expedite—3 POs
Affected: 2 suppliers · 7 MOs · 11 customer ordersSave scenario
04

What-if scenarios

Delay a supplier, lose a machine for a week, accept a rush order. The change is applied to a copy of the plan, and its consequences unfold — orders that slip, delivery performance, the cash low point.

Consequences seen before they are chosen

Predictive · late delivery
Likely arrives five days late
Forecast from the supplier's own receipts
SignalValueWhen
Receipts late4 of 6Last 6 months
Average lateness5 daysPer receipt
Build waiting1 orderBefore forecast
Production and a customer order depend on it
Recommended

Move the expected arrival to the forecast date — previewed, then confirmed.

Plans and promises use the date the goods will really arrive.

05

Suggested actions

Where a risk has a known remedy — move a delivery date, swap a lot, release a reservation, escalate an approval — the engine proposes it beside the risk and records whatever is decided.

From insight to decision in one step

Prediction ledger
Written once, never edited. The score is computed from these rows only.
Frozen atRecordSaidOutcome
02 Sep, 08:14MO-PRED-000182 %Happened
05 Sep, 08:11RM-STEEL-0464 %Averted
09 Sep, 08:15INV-2026-018880 %Happened
14 Sep, 08:10LOT-2026-09785 %Did not
21 Sep, 08:12SO-PRED-000257 %Still open
06

Every prediction on the record

A prediction is written down the moment it is made and never revised afterwards. It is later judged against what actually happened.

An engine accountable for its claims

02

Rehearse the decision before you make it.

Every planner asks the same question: if this changes, what else follows? Because Otiox knows how each record depends on the others, a single change propagates through the plan in seconds — on a copy, never on the real thing.
  1. 01
    Delay a supplier, add demand, take a machine offline, change a lead time
  2. 02
    Every affected manufacturing and customer order, named
  3. 03
    On-time delivery, late orders and cash low point, side by side
  4. 04
    Scenarios saved, shared, and turned into real actions when chosen
app.otiox.com
What-if scenario
Changes apply to a copy of the plan. Nothing real moves until you say so.
IfShenzhen Precisionships+14 daysthen…
MeasureBaselineScenario
On-time customer orders91 %78 %
Manufacturing orders late27
Cash low point€184k€121k
Material to expedite—3 POs
Affected: 2 suppliers · 7 MOs · 11 customer ordersSave scenario
03

From signal to prevented problem

The engine runs every morning, and again whenever something meaningful changes.
  1. 01

    Read the model

    Orders, stock, machines, suppliers and their relationships, as they are now

  2. 02

    Detect

    Fifty detectors look for the patterns that preceded trouble before

  3. 03

    Explain

    Each risk receives a likelihood and the chain of records behind it

  4. 04

    Rehearse

    Test a response on a copy of the plan, or take the suggested action

  5. 05

    Learn

    The outcome is recorded and the prediction is judged

04

A shared view of what lies ahead.

When everyone reads the same risks, ranked the same way, with the same evidence attached, the conversation begins where it should: with what to do, rather than with whether the problem is real.
  1. 01
    Open, acted-on, realised and averted risks in one view
  2. 02
    Filtered by production, purchasing, sales, stock or finance
  3. 03
    An owner for each risk, and the action taken against it
  4. 04
    Risks that close themselves when the underlying records recover
app.otiox.com
Predictive · late delivery
Likely arrives five days late
Forecast from the supplier's own receipts
SignalValueWhen
Receipts late4 of 6Last 6 months
Average lateness5 daysPer receipt
Build waiting1 orderBefore forecast
Production and a customer order depend on it
Recommended

Move the expected arrival to the forecast date — previewed, then confirmed.

Plans and promises use the date the goods will really arrive.

05

The idea, applied.

Scenarios from the demo workspace, walked through step by step — on the same data you receive when you sign up.
01

Late delivery ahead

Which purchase order will arrive late — before it does.
02

A promise the plan can't keep

Promised dates tested against what is possible.
03

Supply shock response

Tariffs, delays, and the response that pays.
06

One ontology, seven ways of seeing it.

Every Command view reads the same things, the same relationships and the same history. An answer found in one becomes evidence in all the others.
01

Accuracy & Calibration

Every prediction frozen and judged against what happened
02

Investigations & Causal Map

Why things happened, with evidence for and against
03

Decisions & Interventions

Who decided, on what evidence, and whether it worked
04

Explorer & Dependency Map

Everything a record is related to, and what is at risk
05

Time Machine

Any record or stock exactly as it was on a date
06

Rules & Controls

Company rules monitored, with attested control runs
07

Questions

01Is the Predictive Engine a black-box AI model?
No. Every prediction comes from named detectors that read your own records and the relationships between them. Each risk shows the records and the path that produced it, so the reasoning can be examined.
02How much history does it need?
Detectors work on current data from the first day — a late purchase order feeding an open manufacturing order is visible immediately. Likelihoods sharpen as Otiox observes how your suppliers, machines and customers actually behave.
03Do what-if scenarios change my real data?
Never. A scenario runs on a copy of the plan. Nothing real changes until an action is chosen and applied, and that action is recorded in the Decision Log.
04How do I know whether the predictions are any good?
Every prediction is frozen when it is made and later scored against what happened. Accuracy & Calibration shows the result for each detector — including those not yet reliable.
Otiox

See what is already in motion.

Open the demo workspace, run the engine, and follow a single risk from the late purchase order to the customer it would have reached.

Get startedLearn more