VYAN, surface by surface.
8 surfaces · Decision Twin · Substrate · EO · Policy · Response · Learning · SAGE · Admin
Eight cooperating surfaces. The Decision Twin is the central object; everything else either feeds it, configures it, or surfaces what it is doing. Open and gated surfaces are clearly marked — the gated ones open in a PULSE workshop or under customer access.
The Decision Twin is the central object.
Other systems build digital twins of factories, products, or networks. VYAN builds a Decision Twin: a continuously-calibrated model of how your enterprise decides — every decision's inputs, its objective function, its constraints, its alternatives, its lineage, its resilience under variability, and the policy it obeys.
Every other surface in VYAN — the Substrate, the Solver, the Decision Policy, the Response Layer, the Learning Loop, SAGE — either feeds the Twin, configures the Twin, or surfaces what the Twin is doing.
Surfaces every decision in the enterprise alongside its math. Each decision with its lineage, alternatives, resilience score, and the policy that governed it.
- ·Lineage · why this decision, traced back to its drivers
- ·Resilience score · in what % of futures does this hold
- ·Alternatives · what else the engine considered
A data model built for how supply actually works.
The Substrate is the part of VYAN you see first when you start a deployment. It is the data model your decisions are made from — organised as 46 dimensions, with typed source dimensions for procurement, production, transport, and inventory. Other planning tools model supply as a single abstraction; VYAN models it as the four distinct decision surfaces it actually has.
The Substrate is versioned. Every change is publishable, reviewable, and rollback-able. This is the layer that makes VYAN auditable to the CIO and defensible to the auditor.
Catalog of reusable substrate templates per industry pattern. A starting point that matches how your enterprise plans, not a blank canvas.
- ·Enterprise Optimization Template · the most complete
- ·Specialized templates · target sub-problems
46 dimensions across calendar, attributes, grains, and measures. Versioned. Publishable. The substrate the Solver runs on. Auditable, reviewable, rollback-able.
- ·46 dimensions · the substrate of decisions
- ·Typed source dimensions · Procurement, Production, Transport, Inventory
- ·259 measures · versioned, publishable
A single dimension card with its attributes panel. The view a data lead uses to govern the Substrate. The Decision dimension carries 29 attributes; the Resource dimension carries 9.
Concurrent multi-horizon, multi-objective optimization. One pass.
Enterprise Optimization (formerly named The Solver) is the math engine. It takes the Substrate and the Decision Policy and produces decisions — not plans. It runs concurrent multi-horizon optimization across all decision classes (procurement, production, transport, inventory, allocation, promising, expediting) against a multi-objective function the buyer authors.
Most planning systems run sequential single-objective solves. Enterprise Optimization runs one pass against all objectives with stated trade-offs. The output is the Pareto frontier — the set of solutions that no other solution dominates. The Decision Policy picks the point on the frontier.
Working capital, service, margin, resilience — each weighted with stated trade-offs. The dollarized objective the Pareto frontier is computed against.
- ·Multi-objective · all four dimensions in one solve
- ·Stated trade-offs · no buried-in-code weights
The set of non-dominated solutions, with the policy choice highlighted. The decision the policy commits to — and the alternatives it weighed.
Plans break. Policies hold.
The Decision Policy is the most important artifact VYAN produces. It is the explicit, governed, versioned statement of how your enterprise decides — not buried in tribal knowledge, not negotiated in meetings, not hidden in spreadsheets. The policy says: under these conditions, decide this way. Under these other conditions, escalate.
Every Decision Policy is authored by humans, governed by Governance Policies, improved by champion-challenger experiments, and audited by SAGE. When the world moves, the policy holds — not because nothing has changed, but because the policy was written to anticipate change.
Author the rules — versioned, reviewable, attributable. A first-class governed artifact, ready for review and Pull-Request approval.
- ·Seven sub-policies · Demand, Supply, Optimization, Autonomy, Publishing, Scenario, Governance
- ·Versioned · every change a Pull Request
Policies in production listed, versioned, and monitored. Each policy with its champion, its challenger, its outcome metrics, and its retirement criteria.
Decisions emit continuously. Between cycles, the world moves.
Most planning systems emit decisions on a cycle — weekly, monthly, quarterly. Between cycles, the world moves and the plan stales. The Response Layer eliminates the cycle. Events are sensed, classified, scored for impact, and responded to within the bounds of the Decision Policy — continuously.
When a response requires a replan, the Response Layer sizes the minimum impact radius. The replan touches only the decisions that materially change. The rest of the plan stays committed. This is what surgical replan means.
Incoming events classified by shock class with proposed responses. A live decision queue — each event scored for EVA-at-risk and routed by policy.
- ·Shock-class classification · Demand Pull-In, Supplier Slip, Capacity Drop…
- ·Per-event EVA-at-risk · dollarized, scored
Surgical replan as the smallest circle of decisions that need to change. The minimum impact replan — the rest of the plan stays committed. No plan thrash.
The system that learns from its own decisions.
VYAN closes the loop. Every decision is logged. Every outcome is measured. Every gap between predicted and realised outcome feeds back into the Decision Policy through champion-challenger experimentation. Master-data drift heals through driver calibration. Policies that no longer hold get retired.
This is what makes VYAN compound over time. Year-one policies are good. Year-three policies are better, because the system has seen three years of how the world actually behaved.
High-level view of how decisions performed against reality. Variance attributed to specific drivers and decisions — feeds the next planning cycle.
Competing policies in production with outcome metrics and retirement criteria. Evidence-based policy promotion. Available in PULSE workshop and the customer-access portal.
Ask the Decision Twin what it is doing. And why.
SAGE is the conversational surface that translates between executives and the Decision Twin. The CFO asks SAGE about working capital. The CSCO asks about expedite freight. The COO asks about supplier risk. SAGE answers from the same Twin, with the same math, with citations to the underlying decisions and policies.
SAGE also remembers. The Institutional Memory captures the why behind decisions — not just the what. When a new planner joins, they inherit the context. When the CFO asks "why did we expedite that order in March," SAGE can answer.
A scripted demo of an executive asking what changed since last review and why. A conversational answer with citations to the underlying decisions and the policy version that governed them.
- ·Schema-bounded translation · SAGE proposes, never auto-commits
- ·Honest "I don't know" pattern · with stated information gaps
Cross-functional view where Sales, Operations, and Finance see the same decision. The same explanation, role-framed per audience. The audit record links back to this conversation.
VYAN runs as real enterprise software.
Admin is what the CIO will want to evaluate during PULSE — not before. Workspaces, Identity, Platform Health, Calendars, Integration, Solver Pool, Feature Flags, Subscription Scope. The names alone signal that VYAN is real enterprise software.
Workspaces · Identity · Platform Health · Calendars · Integration · Solver Pool · Feature Flags · Subscription Scope. The operational surface a CIO and security team review during PULSE. Open in the customer-access portal.
See the platform in your own enterprise.
A PULSE workshop walks the platform against your real decision problem. Two to three hours, no fee, no deck.