This article was originally published on LinkedIn.

The majority of my AI work involves financial analysis and reporting - variance explanations, margin intelligence, anomaly detection, forecasting, audit workflows, and the types of controlled, repeatable processes that CFOs rely on.

This article is different.

This is about the operational side of NetSuite, where inventory, assemblies, constraints, and inbound supply intersect with customer expectations. It's where logistics meets finance, and where a question as simple as "Can we ship 100 units by Friday?" becomes a multi-variable decision problem that NetSuite isn't natively equipped to answer.

You might technically have enough to make it work. For example:

  • 40 units already built
  • Enough components for 30 more units, but trapped inside kits
  • A purchase order arriving Thursday with components for 50 more

While the answer is yes, NetSuite still shows no. And that's because NetSuite only sees static records, not recoverable components, alternative pathways, or conditional strategies.

This is where AI becomes transformational. It becomes an intelligence layer that greatly enhances NetSuite's abilities.

Why NetSuite Struggles With Assembly Reality

Assemblies in NetSuite do a great job of modeling packaging and light manufacturing, but they're not designed for:

  • Dynamic "what-if?" planning
  • Recovering components from assemblies
  • Evaluating cost or margin trade-offs
  • Balancing competing demand
  • Modeling alternative fulfillment paths
  • Understanding constraints
  • Weighing risk

And a large part of that gap is tied to how ATP (Available to Promise) works. ATP is NetSuite's mechanism for determining how many units you can promise to a customer given known supply and demand.

ATP is based on:

  • Inventory on hand
  • Sales orders and allocations
  • Inbound supply (purchase orders, transfer orders, work orders, assembly builds)
  • Expected receipts, depending on setup

But ATP is somewhat limited, because it cannot:

  • Look inside assemblies
  • Recognize components that could be recovered by unbuilding kits
  • Model alternative pathways ("unbuild now vs. wait for inbound")
  • Evaluate margin outcomes
  • Weigh demand priority
  • Simulate build vs. wait vs. reallocate
  • Consider future demand consumption

ATP is deterministic, but real-world planning is not. And that's why teams still resort to using spreadsheets, whiteboards, text messages, Slack threads, and hours of tribal knowledge.

How AI Reshapes the Entire Conversation

AI is great at helping solve multi-variable planning problems that ERPs weren't designed to compute. When given structured NetSuite data - such as assemblies, BOMs, inbound supply, allocations, and demand - AI can generate accurate, explainable scenarios in seconds.

True scenario planning

AI can model several fulfillment paths:

  • Unbuilding assemblies to recover components
  • Leveraging inbound supply
  • Blending recovery + inbound
  • Reallocating components across SKUs
  • Timing-specific build sequences

These are scenarios that would take humans hours of reconciliation and spreadsheet work.

Trade-off evaluation

AI evaluates factors that ATP ignores:

  • Margin differences between scenarios
  • Labor impact
  • Lead-time differences
  • Customer or channel priority
  • Revenue risk
  • Component cross-demand
  • Operational disruption

Instead of a fixed number, you get option sets.

Plain-language recommendations

Instead of scanning shortage reports, you ask:

"What's the maximum number of SKU123 units we can ship this week without jeopardizing other orders?"

AI can return:

  • A precise number
  • The reasoning behind it
  • Risks
  • Assumptions
  • Alternatives
  • The recommended plan

This turns AI into a decision enabler, not just a text generator.

Operational next steps

AI can also propose:

  • Unbuild sequences
  • PO recommendations
  • Build schedules
  • Inventory reallocation plans
  • Timing-sensitive workflows
  • Margin impact summaries

These are the outputs operations leaders have always wanted and NetSuite alone couldn't provide.

Introducing the Assembly Scenario Planner

To make this type of analysis repeatable for real-world NetSuite teams, I designed a prompt that I've been calling the Assembly Scenario Planner. It's essentially a guided operational decision model that evaluates:

  • Assemblies on hand
  • Components trapped inside kits
  • Inbound purchase orders and work orders
  • Current and upcoming demand
  • Margin implications
  • Lead-time constraints
  • Component interdependencies
  • Operational risk

The prompt generates multiple fulfillment scenarios, compares them, and recommends the best path.

What the Planner Produces

  • A comparison of baseline, unbuild, inbound, and hybrid scenarios
  • Detailed explanations of each path
  • Component recovery recommendations
  • Build timing guidance
  • Margin implications
  • Demand protection strategies
  • A single recommended approach
  • Optional: a full HTML report with charts and KPIs

It gives operations leaders the intelligence layer that they need.

The Prompt

Here's the latest version of the Assembly Scenario Planner prompt. 

You are an AI-powered Assembly Scenario Planner for a company running on NetSuite. Use the NetSuite AI Connector to retrieve all required operational data automatically.

-----------------------------
PRIMARY GOAL
-----------------------------
Determine whether we can fulfill a requested quantity of a specific assembly item by a target date, and identify the optimal build/unbuild plan using real NetSuite data.

-----------------------------
USER INPUTS (MINIMAL)
-----------------------------
- Assembly Item (SKU): {{SKU}}
- Target quantity to commit: {{TARGET_QUANTITY}} units
- Target ship date: {{TARGET_DATE}}
- Optional: customer priority rules, margin thresholds, build constraints, risk tolerance.

All other data must be fetched automatically from NetSuite using the NetSuite AI Connector.

-----------------------------
DATA RETRIEVAL REQUIREMENTS
-----------------------------
Using the NetSuite AI Connector, retrieve:

1. **On-Hand Inventory**
   - Current finished goods quantity for {{SKU}}
   - Component quantities for all BOM members

2. **Assemblies & Unbuildable Kits**
   - All assemblies containing components required for {{SKU}}
   - Quantity available for potential unbuild
   - Component yields per unbuild

3. **Inbound Supply**
   - Purchase orders (lines, quantities, expected receipts)
   - Transfer orders inbound
   - Work orders and assembly builds scheduled
   - Any supply expected before {{TARGET_DATE}}

4. **Demand**
   - Open sales orders for {{SKU}}
   - Allocations and reservations
   - Forecasted demand (next 7–14 days)
   - Competing SKUs using the same components

5. **BOM & Cross-Demand Data**
   - Full bill of materials for {{SKU}}
   - All SKUs with overlapping components

6. **ATP Configuration**
   - How ATP is currently calculated
   - What sources of supply/demand it incorporates
   - Any constraints or gaps

If any required field is missing, infer reasonable assumptions and clearly state them.

-----------------------------
SCENARIOS TO MODEL
-----------------------------
Model at least four scenarios:

1) **Baseline – No Unbuild**
   Use on-hand + WIP + inbound supply only.

2) **Scenario A – Aggressive Unbuild**
   Maximize component recovery from all unbuildable assemblies.

3) **Scenario B – Targeted Unbuild + Inbound**
   Selectively unbuild assemblies to balance margin, risk, and operational impact.

4) **Scenario C – Wait for Inbound Only**
   Rely solely on upcoming POs, transfers, and work orders.

For each scenario, calculate:
- Units that can ship by {{TARGET_DATE}}
- Margin impact compared to baseline
- Operational risk level
- Labor implications (if relevant)
- Impact on adjacent or forecast demand
- A short explanation of the path taken

Clearly identify the **recommended scenario** and why.

-----------------------------
OUTPUT: GENERATE A FULL REPORT
-----------------------------
Produce a structured report that includes:

1. Executive Summary
2. Data Snapshot (automatically pulled from NetSuite)
3. Scenario Comparison Table
4. Detailed Evaluation of Each Scenario
5. Recommended Scenario & Reasoning
6. Key Assumptions
7. Risks & Dependencies
8. Final Recommendation

You may also generate:
- Suggested unbuild sequence
- Proposed build schedule
- Suggested inventory reallocation
- Expected margin outcome
- Choke points or constraints to monitor

-----------------------------
OPTIONAL: HTML REPORT
-----------------------------
If the user says "generate full report":
- Return a complete HTML file with inline CSS
- Use Bootstrap + Chart.js
- Include charts: Units by Scenario, Inventory vs Demand, Component Recovery
- Use brand colors: navy (#1b2838) + accent red (#c74634)

-----------------------------
TONE & STYLE
-----------------------------
- Clear, concise, operationally focused
- Write as a decision-support analyst
- Prioritize accuracy, transparency, and actionable recommendations

See the Full Example Report

You can view an example here: Assembly Scenario Planner Example Report

Wrapping Up

This article shows how AI can transform assembly planning in NetSuite, turning guesswork into structured, scenario-driven decision-making. And although most of my work is in the financial realm, the same principles apply beautifully to operational workflows.

What's exciting is that this is just one use case.

In future articles, I'll explore other operational AI workflows for NetSuite, including:

  • Intelligent ATP forecasting
  • Predictive unbuild analysis
  • Build-vs-wait trade-off modeling
  • Multi-SKU component prioritization
  • Operational risk scoring
  • End-to-end assembly planning workflows

Each of these has the potential to unlock a new layer of insight, eliminate hours of manual work, and directly improve financial outcomes.

Stay tuned...