Demand Intelligence
An 8+ model ensemble with intermittency routing, demand sensing and explainable forecasts.
The platform
The planning spine, and everything that plugs into it
Two services compute demand and feasibility once. Scenarios, a control tower, agents and an optimizer all sit on top of the same engines, so analysis, simulation and action share one truth. Switch modules on as you need them.
One Spine
Demand and feasibility, once
Scenario engine
Unlimited branch and compare
Control tower
One exception queue
Agentic AI
Guard-railed, audited
The spine
Two services every module trusts
Instead of bolting features together, iForecastPlan funnels every read through two services. Change an input once and the dependent work recomputes, while the rest stays coherent.
Demand Service
A tier-weighted blend of orders, AI forecasts and external signals. The one answer for how much.
Feasibility Kernel
A capacity check across line, machine, tooling and workforce. The one answer for whether we can.
Dependency-aware propagation
An edit marks only its dependents dirty, so interactive planning stays fast without losing coherence.
One Spine · Demand & Feasibility
single source of truth
Demand planners
Demand Intelligence
S&OP Management
Scenarios · ATP
Buyers & procurement
Materials · MRP
Production planning
Scheduling
Plant & capacity
Line Capacity
Warehouse & inventory
Inventory · MEIO
Quality
Quality · SPC
Finance & leadership
Control Tower
Scenario compare
branch ·
compare · commit
ScenarioOTIF₹/unitMargin
Base94.1%₹18218.2%
Rush order96.2%₹18717.6%
Steel +12%, P3 down93.8%₹20115.1%
Unlimited what-if, on the same spine
A scenario is a copy-on-write overlay, a small set of overrides resolved before the base. Because every engine funnels through the spine, one mechanism makes the whole platform scenario-aware.
- Branch a whole plan cheaply, then compare KPI deltas side by side.
- Commit promotes a branch into the base as an immutable snapshot.
- Model a promo, a launch, a tariff or CO2e as a lever, with no new engine.
One control tower, every exception
Worklist items, SPC and maintenance alerts, stock warnings, Item 360 flags, OTIF risk and supplier blast-radius all land in a single, role-routed queue.
- Each exception carries severity, an owner role and a one-click simulate fix.
- Push delivery that suits the floor, including email and WhatsApp for criticals.
- Closure tracking, so nothing slips between shifts.
Exception queue3
critical
- StockCirclip 8mm below reorder, Kanpur₹5.0L risk
- SPCCpk drift on Line 4Cpk 1.12
- SupplierV123 PO slip, 3 days lateOTIF -1.4
- OTIFOrder #4471 at risk Thu
Agent proposalguardrails passed
Trigger: rush order #4471 would miss Thursday.
Proposed deltas, in a scenario
Add 1 OT shift on GRIND, pull V128 for circlips
+2.1 pts
OTIF, 0.6 pt margin cost
Agents that act, with a planner in the loop
Your engines have clean signatures, which makes a ready tool surface for agents. An agent senses an exception, builds a candidate scenario, and proposes an action with the money impact attached.
- Guardrails first. Every proposal passes margin, capacity and compliance checks before it can commit.
- A full audit trail of trigger, assumptions, tool calls, approver and outcome.
- Exposed over MCP, so your own copilots can call engines through one contract.
Optional optimizer that adds to your heuristics
An optimization layer runs inside scenarios, warm-started by your heuristics. Start with sourcing and production allocation that minimizes total cost under finite capacity, vendor quotas, MOQ and lead times.
- Heuristics stay as the fast path and fallback, so the UX stays snappy.
- Explainable output. Which constraint bound, and the shadow price behind it.
Allocation result
Vendor A
58%
Vendor B
30%
Vendor C
12%
Binding: GRIND capacity. Shadow price ₹1,240 per shift.
Modules
Eight modules, one data model
Explore the workspace, then dive into any module below.
Modules
AI forecast against actuals, with MAPE, intermittency routing and seasonal-shift sensing.
Planning & Network
Multi-level supply planning across a 6-layer MRP network with WIP deduction.
Production Scheduling
A constraint-aware Gantt with sequence optimization and changeover-minimizing runs.
Materials & MRP
A network cascade with netting and Clear-to-Build readiness down to the component.
Inventory & MEIO
Quantile safety stock, reorder and multi-echelon placement across the network.
Capacity
A finite capacity ledger with utilization and contention across work centres.
Quality & SPC
Statistical process control, Cpk drift detection and predictive maintenance.
Order Promising
Available and capable-to-promise, with an OTIF impact simulator for every yes.
Item 360
One screen that brings 10+ engines together for any SKU across the business.
Trust
Governance built in, not bolted on
Role-based access
Every exception and action routes to the right planning role.
Full audit trail
Feasibility, promise and agent-action logs for every decision.
Single-tenant
A dedicated, isolated environment, encrypted in transit and at rest.
Open by contract
A REST API, webhooks and an MCP server over the engine surface.