Narrative Insights: AI-Generated Summaries for Standard Reports and Records

What Changed

Narrative Insights is a generative-AI feature that renders a summary dialog when a user clicks Generate Insight on a supported report or record. Key details:

• Scope: UI-only. There is no SuiteScript module, no REST endpoint, and no SOAP operation for Narrative Insights. You cannot programmatically trigger or retrieve insights.
• Preference: Controlled by Enable Narrative Insights on the AI Preferences page (Setup > Company > AI Preferences). Enabled by default. Only Administrator-role users can toggle it.
• Availability: All NetSuite regions and all system-supported languages.
• Output variability: Summaries are non-deterministic — the same underlying data can produce different text on each generation. Oracle's own disclaimer warns the output may contain errors.

Supported Reports

Only standard reports are supported. Customized reports are explicitly excluded. The supported report areas are:

• Banking/Budgeting: Budget vs. Actual
• Customers/Receivables: A/R Aging Summary, Customers by Sales Rep Summary
• Financial: Income Statement, Comparative Income Statement, Balance Sheet, Comparative Balance Sheet, Cash Flow Statement, Trial Balance
• Inventory/Items: Inventory Activity Detail, Inventory Back Order Report, Stock Ledger
• Marketing: Campaign ROI Analysis Summary & Detail
• Order Management: Open Return Authorizations, Return Authorizations Register, Open Sales Orders, Sales Order Register
• Payroll: Payroll Summary, Payroll Liabilities
• Pipeline Analysis: Sales Activity by Sales Rep (Summary & Detail), Total Open Opportunities Detail, Total Open Estimates, Opportunities Won, Pipeline by Sales Rep Summary, Pipeline by Status Summary
• Purchases: Open Purchase Orders
• Sales: Sales by Item/Customer/Sales Rep (Summary & Detail)
• Sales Orders: Sales Orders by Item/Customer/Sales Rep (Summary & Detail)
• Vendors/Payables: A/P Aging Summary

Supported Records

• Accounting: Inventory Items
• Customers: Credit Memos, Customer Payments
• Financial: Journal Entries
• Relationships: Customer 360
• Sales: Invoices, Sales Orders, Estimates
• Support: Cases

Data-Size Limitations

The AI generation can fail if the report or record contains too much data or the generation times out. Oracle recommends shortening the date range, applying filters (by customer, subsidiary, etc.), or reducing line counts on records.

What This Does NOT Affect

This is important for developers and integrators to understand:

• No new SuiteScript 2.x modules, no N/ai or similar namespace exists for this feature.
• No new fields are persisted on any record — insights are ephemeral, generated on-the-fly in the UI.
• No REST (/services/rest/record/v1/...) or SOAP/SuiteTalk operations are added.
• Saved searches, custom reports, and SuiteAnalytics Workbook reports are not supported.
• No workflow or SuiteFlow trigger fires on insight generation.

What to Do

1. Admins: Decide whether to leave the Enable Narrative Insights preference on (default) or disable it. This is a company-wide toggle — there is no per-role or per-user control mentioned in the documentation.
2. Developers/Integrators: No code changes are required. The feature has no API surface. If you were hoping to consume AI-generated summaries programmatically, this feature does not offer that capability.
3. Customization owners: If your workflows rely on customized versions of the supported reports, note that Narrative Insights will not be available on those customized variants — only the out-of-the-box standard versions are supported.
4. Security/compliance teams: Review Oracle's LLM model mapping (SuiteAnswers article referenced as LLM Mapping for Generative AI Features, requires NetSuite login) to understand which AI models process your account data. The documentation does not specify whether data leaves the NetSuite environment or which LLM provider is used.

Open Questions

The source documentation is light on technical specifics. Key unknowns include:

• Which release introduced this feature (no version is stated in the documentation).
• Whether insight generation consumes governance units or counts toward any usage quota.
• Whether the AI processes data within NetSuite's infrastructure or sends it to an external LLM provider.
• Whether per-role or per-user access control is possible beyond the single company-wide preference.