aPlanner is in MVP validation (Oct–Dec 2026). Figures are design targets. Join the 2027 Korea pilot →

Home
/
Platform

THE APLANNER PLATFORM

Four AI modules, one planning engine

Forecasting tells the fab what the market will want; scheduling turns that into a feasible, yield-aware plan; simulation lets planners test every decision; analytics keeps everyone on the same live numbers.

aPlanner is the first product of the Amoza AI platform — from shop floor to customer door. Next: aMonitor (factory monitoring).

Join the 2027 Korea pilot program Prototype preview

aPlanner MVP — recommendations panel with impact one-liners — Recommendations — HIGH cards with impact one-liners, Preview and Explain

Four intelligent modules. One planning engine.

Each module replaces a spreadsheet, not a dashboard on top of one. All outcome figures are design targets. Validated on synthetic NAND-flash datasets — pilot results coming 2027.

AI Demand Forecasting

Ensemble ML per bit-density tier / SKU with confidence intervals. Target: 85–90% accuracy.

Intelligent Scheduling

Constraint-based, yield-aware optimizer over 50–100+ steps with automatic re-planning. Target: full-fab schedule in < 30 min.

Scenario Simulation

Monte-Carlo what-if engine with templates and side-by-side KPI comparison. Target: minutes per scenario.

Real-Time Analytics

Live planning KPIs, anomaly detection, explainable recommendations, Korean / English views. One source of truth.

Target figures are design targets from simulation on synthetic fab data. aPlanner has no customer deployments yet; first pilot results are expected in 2027.

WHAT CHANGES FOR THE PLANNER

From a weekly batch plan to a living plan

Today: an APS run plus Excel that is obsolete by Wednesday, one or two scenarios per decision, yield and qualification handled in the planner's head, fifteen e-mails per change.

With aPlanner: a plan re-planned incrementally to the last event, three ranked feasible options in under a minute, a yield-aware constraint model, and one commit that fans out automatically.

What changes for the planner: today versus with aPlanner — What changes for the planner — weekly batch plan → living plan; 1–2 scenarios → 3 ranked options; 15 e-mails → one commit

MODULE 1 · AI FORECASTING

Predict demand with confidence intervals

What it is

Ensemble ML (LSTM, Prophet, XGBoost, Transformer) forecasting demand per bit-density tier / SKU, 1 day to 24 months, with confidence intervals; the best model per segment is chosen automatically and retrained as data arrives.

What problem it solves

Excel forecasts reach 55–65% accuracy, worth US$20–50M a year of over- / under-production per large fab; AI / data-center orders swing 30%+ in a quarter.

What benefit it brings

Target 85–90% accuracy (+20–30 pts); less inventory in the wrong place; demand sensing catches hyperscaler swings weeks earlier.

Target: 85–90%forecast accuracy vs. 55–65% with Excel

Explore forecasting

UC-7 user flow — quarterly demand revision and node-migration scenario — UC-7 — Quarterly demand revision and node-migration scenario (NAND)

MODULE 2 · INTELLIGENT SCHEDULING

A feasible, yield-aware plan in minutes

What it is

Constraint-based optimizer across tools, materials, labor, maintenance and changeovers, yield-aware over 50–100+ wafer steps, with automatic re-planning on disruptions.

What problem it solves

3–5-day manual cycles; about 70% of schedules need manual fixes; plans are outdated before they are finished; knowledge lives in planners’ heads.

What benefit it brings

Target: full-fab schedule in under 30 minutes (80% faster); 99%+ feasible plans; bin-sort allocation maximizing revenue per wafer.

Target: < 30 minfor a full-fab schedule, 99%+ feasible

Explore scheduling

aPlanner MVP — compare candidate plans side by side — Compare — finish date, peak tool load, over-capacity, all-safe check

MODULE 3 · SCENARIO SIMULATION

Test every decision before you make it

What it is

Monte-Carlo what-if engine with templates (tool down, +20% hyperscaler order, yield −10%, node migration), side-by-side KPI comparison and an AI-recommended plan.

What problem it solves

A scenario takes days in Excel, so planners test 1–2 options and decide by gut feel — an estimated US$15–25M a year of sub-optimal decisions per large fab.

What benefit it brings

Target: minutes per scenario (95% faster); unlimited scenarios; quantified risk; decisions the whole team can defend.

Target: minutesper scenario instead of days

Explore simulation

aPlanner MVP — AI assistant with two re-plan options — AI Assistant — “hold Customer K at +2 d without hurting the automotive lot” → 2 options

MODULE 4 · REAL-TIME ANALYTICS

One live source of truth for the whole fab

What it is

Live planning KPIs (adherence, OTD, WIP, utilization), anomaly detection, explainable recommendations; Korean / English planner and executive views.

What problem it solves

Data siloed in MES, ERP and spreadsheets; planners spend about 70% of their time gathering it; management sees yesterday’s numbers.

What benefit it brings

One always-current source of truth; target: planner time on strategy 20% → 60%; every recommendation auditable.

Target: 20% → 60%of planner time on strategy

Explore analytics

aPlanner MVP — executive analytics dashboard — Analytics — executive view: adherence, OTD, WIP, utilization

INNOVATION & IP

What is new and patentable

Three things that today live in three different heads — yield, constraints and scenarios — planned together. KIPO patent applications planned Dec 2026 and Q1 2027.

Yield-aware constraint scheduling

The optimizer takes live yield and bin distributions per process step as inputs, so wafer starts, expected good die and bin mix are planned together across 50–100+ steps — instead of scheduling starts and hoping the yield follows.

KIPO patent application #1 — planned Dec 2026

Node-migration planning

Models learning curves, ramp rates and capacity shifts when a line moves 176L → 236L → 300L+, planning old and new nodes as one system with shared tools and demand.

KIPO patent application #2 — planned Q1 2027

Explainable re-planning

Every schedule change carries the triggering event, the constraints it hit and the KPI delta; frozen / flex / far zones bound how much the plan may move; the planner approves, edits or rejects in Korean or English.

Korean UI + planner-approval workflow (in the MVP)

Read about the three innovations

ARCHITECTURE

Fits the fab's existing stack — data never leaves the site

MES · ERP (SAP) · yield / bin DB · planner spreadsheets → data layer (aFlow · aConnect) → AI core (aCore) → planner workspace (KR / EN, approvals) → back to MES / ERP.

aPlanner service logic: fab data sources → data layer → AI core → planner → back to MES / ERP — Service logic — MES · ERP · yield / bin DB · spreadsheets → data layer (aFlow · aConnect) → AI core → planner workspace → back to MES / ERP

Deployment

On-premise or private cloud inside the fab network (air-gapped option); SaaS on AWS Seoul for suppliers and SMEs; Kubernetes, same code everywhere.

Integration

Pre-built connectors for in-house MES, SAP ERP, yield / bin databases and Excel imports; read-only first, write-back after sign-off.

Security & governance

RBAC, SSO, full audit trail; planner-approval workflow. Roadmap: SOC 2 Type II (planned 2027–28), ISO 27001 (2028), K-ISMS.

Architecture & integration details

Technology

Backend	Python / FastAPI
Data	PostgreSQL + TimescaleDB
Optimization	OR-Tools / MILP (CP-SAT)
AI / ML	PyTorch, XGBoost
Frontend	React (KR / EN)
Runtime	Kubernetes on-prem or AWS Seoul

Security & deployment

On-premise / private cloud first for Korean fabs; air-gapped option

SaaS on the AWS Seoul region for suppliers and SMEs

Encryption at rest and in transit; RBAC; SSO; audit trail

Planner-approval workflow — AI recommends, people decide

Roadmap: SOC 2 Type II (planned 2027–28) · ISO 27001 (2028) · K-ISMS (planned)

Security & deployment details

DEVELOPMENT STAGE

Pre-commercial MVP, in validation

Not commercialized; no revenue; no customers yet. Working prototypes of scheduling, simulation and analytics on synthetic fab data; forecasting engine prototyped, dedicated view in the December 2026 build. The first Korean fab pilot is planned for 2027.

Done

Phase 1 — Foundation

Apr–Jun 2026

Done

Phase 2 — Concept

Jul–Sep 2026

Working prototypes on synthetic fab data

Now

Phase 3 — Validate

Oct–Dec 2026

MVP validation, Korean UI

Planned

Phase 4 — Launch prep

Jan–Mar 2027

aPlanner MVP — Monday daily digest over the Gantt — Daily Digest — 3 weekend events, 2 folded into a proposed re-plan, 1 needs the planner’s call

aPlanner MVP — commit confirmation dialog — Commit confirmation — what changes and who gets told, with Explain before signing

EDITIONS & PRICING · KRW FIRST

Start with a pilot, grow by edition

Planning Readiness Assessment → pilot on one line (KRW 75–150M, 8–12 weeks) → Starter subscription from KRW 120–300M per year (≈ US$100–250K). One price list on both Amoza sites; indicative until the first Korean pilot is contracted.

View the price list

Be the first Korean fab to run aPlanner

The 2027 Korea pilot program: a paid readiness assessment, an 8–12-week pilot on one line, and a founder who sits next to your planners.

Join the 2027 Korea pilot program Talk to the founder

No commitment

Prototype walkthrough

Planning readiness assessment