If someone on your team asks for an AR aging snapshot, a quick data quality check, or a list of who has admin access — how long does it take to get that answer? For most businesses running NetSuite as their primary operational system, the honest answer is: longer than it should. By the time someone logs in, builds a saved search, runs it, formats it, and sends it — you’ve got 10 new emails you could have responded to. More often than not, the person who needs the information just ends up waiting for someone in accounting to pull it.

We’ve been testing a different approach: connecting AI directly to NetSuite via the MCP (Model Context Protocol) connector SuiteApp and running queries in plain English. No saved searches, no logging into the UI. Just ask the question and get the answer.

This post is the anchor for a series where we document exactly what that looks like in practice — what works, what doesn’t, and where you need to be careful.

What the NetSuite MCP Connector Actually Does

The connector gives an AI direct access to your NetSuite account through SuiteQL — NetSuite’s SQL-based query language. For our purposes, we used Anthropic’s Claude.ai. One reason we chose Claude specifically: on Team accounts, Anthropic does not use your data to train its models, which matters when you’re querying live financial and customer data.

In practical terms, the AI can:

  • Pull any record type and read all accessible data on it
  • Run custom SuiteQL queries against any accessible table — without you needing to understand SuiteQL
  • Return structured results you can read, summarize, or act on immediately — including executive-ready visualizations
  • Create or edit records directly in your system (more on why this requires careful controls below)

What it cannot do is just as important to understand. It can’t trigger workflows or run SuiteScript. It can’t perform UI-driven operations like merging records or running mass updates. For most of the use cases in this series, you won’t need write access at all — and we’d recommend restricting it to read-only for the majority of AI-assisted workflows.

Who This Is For

This series is most relevant if you’re in finance ops, a NetSuite admin, or leadership and you want faster answers without spinning up a developer or building another saved search.

It’s not a replacement for formal reporting, audit trails, or your period-end close process. Think of it as a first-pass investigative tool — fast, flexible, and good at surfacing things worth a closer look.

One Thing Worth Understanding Before You Start

Everything the AI can access comes down to what permissions you give it. That’s the part that’s essential to keep in mind: incomplete permissions can look like missing data — the query runs cleanly but returns zero rows because the role doesn’t have access, not because the data doesn’t exist.

Before you trust any output, test it against something you already know the answer to. If the results match your expectations, you’re in good shape. If they don’t, investigate the role permissions before going further. We document this tension throughout the series — but it’s worth knowing upfront.

What We’ve Tested

We’ve been running live queries against SuiteCentric’s own NetSuite account and documenting what we find. Here’s the full series:

How to Get Your NetSuite AR Aging Instantly Without Running a Report — Finance teams check AR aging constantly. We queried open invoices for a customer, bucketed them by current / 30 / 60 / 90+ days past due, and had a full aging breakdown in seconds. We also verified that every overdue invoice had actually been sent — a simple step that often reveals a communication breakdown before it becomes a collections problem.

Find NetSuite Data Quality Problems Fast with AI — We ran a set of queries looking for test records, duplicates, and missing required fields. The results were more substantial than expected — and more actionable than a spreadsheet export.

We Tested AI for Post-Meeting NetSuite Updates. Here’s the Reality. — We used Claude to update opportunities, add contacts from meeting transcripts, and log contact-level next actions directly in NetSuite. Includes a close look at the confirmation workflow you need in place before giving AI write access to live data.

I Asked AI to Pull Our Customer Revenue Snapshot from NetSuite — Here’s What Came Back — We queried invoiced amounts by customer year-over-year and layered in actual cash collected to separate billing activity from collections reality. The result: a two-section snapshot — top performers and accounts to watch — that makes managed services account reviews dramatically faster.

Using AI to Integrate Systems (Almost) — We tested using Claude and the NetSuite MCP connector to move Shopify orders into NetSuite instead of building a custom integration. It’s not a real integration — no live sync, no monitoring — but a repeatable manual process fast enough to feel like one, with some real lessons on where AI held the line and where its own fix got messy.

Auditing NetSuite User Access and Roles with a Simple Query — We asked for every full-license role holder in the account, flagged who carried admin access, and pulled login history for each person — no query logic written by hand. The login history piece needed a permission grant to fully resolve, and the audit caught an account with an unusual internal ID that nearly got excluded despite having real admin rights.

Curious what this could look like for your team? We’re happy to walk through a live demo against your NetSuite environment. Book a free consultation to talk through what makes sense for your setup.