NetSuite context intelligence · Preflight
Point SuiteContext at a field, script, or workflow in an account you didn’t build — and get a deterministic verdict on what depends on it, what conflicts, and what to check before you deploy.
Read-only · Deterministic — no LLM · Never trained on your data
Exhibit A · what a preflight looks like
Two scripts writing one field on the same trigger — found by static analysis across every active deployment.
Confirmed, observed, inferred, unknown-risk. What we know, separated from what we can’t see.
Preflight report
№ 0042 · read-only
Proposed change: repurpose custbody_priority_code on Sales Order
UE_set_priority.js and SL_order_sync.js both write this field on afterSubmit. Execution order unmanaged — last writer wins.
3 saved searches filter on this field; one feeds the nightly revenue report a controller reads every morning.
An active workflow transitions on the same record type; user-event scripts complete before its afterSubmit actions — verify the state it expects.
Not readable via any API. Confirm on the Scripted Records page before deploying.
1 blocking · 2 checks · same input, same verdict — every run
When something isn’t readable via any API, the report says so — and tells you exactly where to look.
No model in the path. Same input, same verdict, every run — showable to a client.
“Who changed this, and why?”
System notes say what and when — never why, and never across the whole account at once.
“What breaks if I touch this field?”
There’s no native view of what depends on what. You find out in production, after the deploy.
“How do I onboard onto this?”
Mature accounts are tribal knowledge nobody wrote down. The person who built them left.
The catch nothing else surfaces
In most accounts, script execution order is unmanaged and invisible — grouping rules silently override naive expectations, and the configured order isn’t readable via any API. So when two user-event scripts write the same field on the same trigger, last writer wins — and nobody chose the winner.
UE_set_priority.js · afterSubmit
rec.setValue('custbody_priority_code', score)
SL_order_sync.js · afterSubmit
rec.setValue('custbody_priority_code', extVal)
■ Both deployments active on Sales Order. Execution order unmanaged — last writer wins.
SuiteContext statically analyzes every script source against every active deployment — cross-script write collisions joined on record type and entry point. Not something you can eyeball across a few hundred files.
These conflicts don’t error. They produce intermittent wrong values that pass UAT, then surface in month-end reports. Teams burn days on them because no tool points at the collision.
You get the colliding scripts, the field, the trigger, and what to check — cited and repeatable. Field explorers show IDs. SuiteContext explains consequences.
The honesty model
No tool can know everything about a NetSuite account, and the ones that pretend to are the ones that get you. Every SuiteContext finding carries one of four grades. That’s the difference between a senior consultant and a confident guess.
Read directly from the account: script source, deployments, field definitions, system notes.
Seen in structure or history — present and real, but its runtime behavior isn’t guaranteed.
Follows from how NetSuite executes — entry-point grouping, script/workflow ordering rules — applied to your inventory.
A landmine no API exposes — like the configured Scripted Records order. We flag it and tell you exactly where to look.
The AI era
Oracle’s free MCP, Copilot, and Cursor made agents able to read and change NetSuite — which makes a neutral, deterministic check more necessary, not less. Ask an LLM “what breaks if I change this field” and it improvises from raw rows, differently every run. We ride the MCP; we don’t compete with it.
Access — what the official MCP returns
> query: scripts on salesorder
rows: 47 · scriptdeployment · …
rows: custbody_priority_code · …
… 312 more rows. Your AI improvises from here.
Judgment — what SuiteContext returns
Conflict: 2 scripts write custbody_priority_code on afterSubmit.
Confirmed from source · execution order unmanaged — last writer wins · check the Scripted Records page before deploying.
Deterministic. Cited. The same answer every time.
How it works
The same read-only role + token setup every integration vendor asks for — no install in your account, nothing in the browser’s data path.
Step 1 · Connect
Create a read-only role and token over NetSuite’s supported APIs (SuiteQL, SDF). Metadata and script structure — record values stay opt-in. Revoke in one click.
Step 2 · Analyze
Fixed rules, static analysis of script source, and encoded tribal knowledge — entry-point grouping, script/workflow interaction — run server-side. No LLM in the path.
Step 3 · Preflight
Dependencies, conflicts, risk flags, and what to check — graded Confirmed / Observed / Inferred / Unknown-risk. Export it as grounded, cited context for your own AI.
Why ERP teams can trust it
Everyone wants AI help with NetSuite. Nobody wants their ERP feeding a model. So the analysis is deterministic TypeScript end to end — and AI is something you opt into, with context that’s grounded and cited, never trained on.
Fixed rules, not a model. Same input, same verdict.
Not for training, not for “improvement,” not ever.
Field IDs, types, script structure. Record values are opt-in; PII redacted.
A role you create and revoke. Nothing installed in your account.
Subscription key — stop paying, revoke, and it all goes dark.
Confirmed, observed, inferred, unknown-risk.
Who it’s for
“I just got a client account and need to understand it fast — without burning unbilled hours.”
Understand an unfamiliar account in the first session. Scope and quote engagements on evidence, and hand the client a risk map that sells the next phase.
“The last admin left. Nobody knows what half of this does.”
The documentation the last admin never wrote — grounded in what’s actually in the account, not what anyone remembers.
“I need to change this script and I don’t know what I’ll break.”
Impact and conflict analysis before the deploy — plus grounded, cited context to hand your AI before it writes a line.
The honest map
We position around the job, not against anyone. Here’s what each alternative is genuinely good at — and the job SuiteContext exists for.
| Alternative | Genuinely good at | Where SuiteContext fits |
|---|---|---|
| Governance suites (Strongpoint, Salto)Compliance, change management, full documentation. | Compliance, change management, full documentation. | Different job: fast, self-serve preflight in minutes — at mid-market price, not a multi-week rollout. |
| Free field explorers & FlashlightField IDs, hidden fields, free dependency visibility. | Field IDs, hidden fields, free dependency visibility. | They depict; we interpret. Which scripts collide, what’s risky, what to check — graded by confidence. |
| Official NetSuite MCP · Copilot · CursorGiving your AI access, and generating SuiteScript fast. | Giving your AI access, and generating SuiteScript fast. | We supply the judgment layer they don’t: deterministic, cited verdicts about this account. We ride the MCP. |
| By hand (the current default)Free, fully flexible. | Free, fully flexible. | Stop reverse-engineering accounts by hand. Minutes instead of days, and it catches what a human skims past. |
Start here
We’re pre-launch and building with design partners. Two ways in — one available this week, one that locks your price for life.
The founding teardown
Pick the field, script, or workflow you’re most afraid to touch. We hand-run the full Preflight on it and walk you through the findings live.
Founding access · first 50 teams
The automated product, as it ships to design partners: connect a read-only token, get the inventory, dependency maps, conflict catches, and risk audit in minutes.
Indicative founding price — final plans set with design partners. No launch date promised; access and influence are.
Why I’m building this
“I’m a software engineer who’s spent five-plus years deep inside NetSuite — not just administering it, but understanding why it’s built the way it is. I’ve lost too many days reverse-engineering accounts I didn’t build. SuiteContext is the tool I kept wishing existed, and the teardown is me proving it on your account before asking you to believe anything.”
FAQ
A read-only role and token over NetSuite’s own supported APIs (SuiteQL, SDF) — the same ~30-minute checklist Celigo or Boomi asks for. No install in your account. Metadata and script structure, not record values; revoke in one click. For a teardown, an SDF export or files you send work too.
No. The analysis is deterministic TypeScript — there is no model in the path, so your audit data is never sent to an AI because it isn’t sent anywhere to be analyzed. AI explanations are opt-in, off by default, grounded in the engine’s findings, cited, and never trained on your data.
Use it — it’s great free documentation and dependency visibility. Flashlight depicts; SuiteContext interprets: which scripts collide, what’s risky, what to check, graded by confidence. Different job.
Exactly — and we ride it. The MCP gives your AI access: raw records and SuiteQL. It doesn’t give it judgment: ask “what breaks if I change this field” and it improvises from raw rows, differently every run. We return a deterministic, cited verdict — the same answer every time.
Great for one script. Account-level questions — cross-script write collisions across every deployment — need static analysis, and an LLM’s answer changes every run. Ours is repeatable and showable to a client.
No — and we don’t pretend to. Findings are labeled Confirmed, Observed, Inferred, or Unknown-risk, and the report tells you exactly where to verify the things no API exposes. That honesty is the point.
We’re building with founding design partners and promise access and influence, not a date. Teardowns run now at $300. Founding subscriptions start at $99/mo per account (unlimited users), grandfathered for life for the first 50 teams.
Bring one scary customization. We’ll run the Preflight by hand and show you exactly what SuiteContext would catch — for $300, credited if you stay.