Last week I shared a system prompt that turns AI into a senior SuiteScript developer. This week, it gets a counterpart: a Senior NetSuite Finance Functional Consultant that handles the other half of the equation - things like configuration, accounting guidance, process design, and troubleshooting.
This new prompt uses the same approach as last week\'s prompt, but provides a different area of expertise. And like last week\'s prompt, this one is also being provided for free.
Why a Finance Consultant?
The SuiteScript developer prompt solved the code problem. But most NetSuite questions aren\'t code questions.
Instead, they\'re questions like: How do I configure three-way matching? Why doesn\'t my consolidated balance sheet balance? What\'s the best way to handle revenue recognition for bundled SaaS contracts? Should I use a workflow or a script for vendor bill approval?
When I asked AI these questions, I kept getting:
- Generic accounting advice with no NetSuite-specific configuration steps
- Fabricated navigation paths that don\'t exist
- Custom code recommendations for problems native features solve in minutes
- No mention of GL impact, prerequisites, or downstream effects
- Zero push-back when I asked for something that would create audit risk
So I built another system prompt. One that thinks like a CPA who\'s done many NetSuite implementations - not a generic chatbot that read a NetSuite help article once.
What This Prompt Does
Load it as a system prompt. Ask any NetSuite finance question. The AI responds as a senior finance consultant with four built-in behaviors:
Native-First Decision Framework - Before recommending customization, the AI evaluates native configuration, SuiteFlow workflows, saved searches, Financial Report Builder, and SuiteApps. If a workflow handles vendor bill approval, it says so - and explains why SuiteScript isn\'t needed.
Configuration with Context - Every recommendation includes exact navigation paths, prerequisite features to enable, GL impact, common pitfalls, and downstream effects on reporting and integrations. Not just "turn on this feature" - the full picture.
Structured Response Templates - Three templates that activate automatically based on your question type: Configuration (UNDERSTAND → PREREQUISITES → CONFIGURE → VERIFY → DOWNSTREAM IMPACTS), Advisory (UNDERSTAND → ASSESS → EVALUATE → RECOMMEND → IMPLEMENTATION PATH), and Troubleshooting (DIAGNOSE → INVESTIGATE → FIX → PREVENT).
Compliance-First Priority - Data integrity and internal controls are Priority 1. If you ask it to remove journal entry approvals, it pushes back - explains the SOX, SoD, and audit trail implications - and offers five alternatives that reduce friction without eliminating controls.
Real Examples
Example 1: Vendor Bill Approval - It Recommended a Workflow
The ask: "We need approval routing for vendor bills over $10,000. Bills under $10K should auto-approve."
The AI evaluated native approval routing (insufficient for threshold-based auto-approval), then recommended a SuiteFlow workflow - not SuiteScript. It provided a complete workflow design: four states (Auto-Approved, Pending Approval, Approved, Rejected), transition conditions, escalation rules (48-hour reminder, 96-hour escalation), and the exact navigation paths to build it.
It also included a comparison table showing why SuiteFlow beats both native approval routing and custom SuiteScript for this specific requirement.
Example 2: Three-Way Matching - Full Configuration Walkthrough
The ask: "How do I configure three-way matching for our AP process? We receive about 200 vendor bills per month."
The response followed the Configuration template exactly:
- Prerequisites - Six items to enable, each with exact navigation paths
- Configuration - Six steps including tolerance settings with a recommended-values table
- GL Impact - Full entry flow across all three stages (PO → Item Receipt → Vendor Bill), including Purchase Price Variance treatment
- Verification - Eight test scenarios plus a table of seven common pitfalls with causes and prevention
- Downstream Impacts - Eight areas affected (reporting, month-end close, AP workflow, vendor relationships, payment runs, integrations, audit trail, PPV)
Example 3: Month-End Close Optimization
The ask: "Our month-end close takes 15 business days. We\'re a wholesale distributor on OneWorld with 2 subsidiaries. Help us get to 5 days."
The AI delivered a complete close optimization plan: current-state assessment, native automation recommendations for each bottleneck (inventory reconciliation, intercompany entries, bank reconciliation), a 5-day close calendar with pre-close tasks, and a three-phase rollout plan targeting 10 days → 7 days → 5 days over six months. All native features, no SuiteScript, with distribution-specific considerations throughout.
Example 4: Audit Risk - Pushed Back with Alternatives
The ask: "Can we remove the approval workflow on journal entries? It slows us down."
The AI pushed back diplomatically. It explained four specific compliance risks (SOX exposure, segregation of duties, audit trail loss, error prevention), then offered five alternatives: threshold-based auto-approval, role-based auto-approval, category-based routing, streamlined approval with backup approvers, and a batch approval dashboard. It then asked clarifying questions about what specifically is slow - because the real bottleneck is rarely "approvals exist" and usually "approvals are routed poorly."
Example 5: Code Request - Redirected with a Functional Spec
The ask: "Write me a SuiteScript that automatically creates journal entries when invoices are approved."
Instead of writing code (which is outside this prompt\'s scope), it provided a complete functional specification that a developer can implement: trigger event, seven business rules in plain language, input/output field tables, two worked GL impact examples with exact debit/credit entries, and an error handling matrix covering seven failure scenarios. Then it offered to help with the accounting and configuration side.
That\'s the division of labor. The Finance Consultant prompt handles the "what" and "why." The SuiteScript Developer prompt handles the "how."
How I Tested It
I used the prompt with 25 test scenarios:
- 44% were configuration questions (COA design, AP, ARM, multi-book, multi-currency, tax, fixed assets)
- 32% were advisory questions (close optimization, implementation strategy, board reporting, COA design)
- 16% involved troubleshooting (balance sheet discrepancies, duplicate rev rec, cash flow issues)
- 20% were edge cases (ambiguous input, code requests, audit risk, uncertainty, conflicting requirements)
Two Prompts, One Team
Here\'s how the two prompts work together:
- Finance Consultant - "Should we use a workflow or a script for vendor bill approval?" → Recommends workflow, designs the approval logic, defines thresholds, maps the GL impact
- SuiteScript Developer - "Build the workflow action script that sends a Slack notification when bills over $50K are approved" → Writes production-ready code with error handling, governance awareness, and deployment instructions
The Finance Consultant decides what to build and why. The SuiteScript Developer builds it. Just like a real implementation team.
Two Versions - Full & Lite
Full Version (~3,500 tokens) - Best for complex advisory, implementation strategy, and multi-turn consulting conversations. Includes 13 financial modules, 6 industry verticals, implementation methodology, and integration patterns.
Lite Version (~1,500 tokens) - Best for quick configuration questions and smaller models. Preserves core behaviors: priority hierarchy, native-first framework, response templates, edge case handling.
How to Get Started
Step 1: Download the prompt.
Step 2: Paste it into your AI tool\'s system prompt. Claude, ChatGPT, Gemini — they all work.
Step 3: Try this first:
We\'re a SaaS company on OneWorld with US and UK subsidiaries. We need to configure ASC 606 revenue recognition for contracts that bundle annual software licenses with implementation services. About 200 new contracts per quarter. What\'s the approach?
You\'ll get a structured response with ARM configuration steps, performance obligation setup, SSP allocation guidance, exact navigation paths, GL impact, and common pitfalls - all tailored to SaaS.
Tips for Best Results
- Mention your industry. SaaS, distribution, manufacturing, and professional services have fundamentally different accounting patterns.
- Specify your subsidiaries. OneWorld vs. single-subsidiary changes many recommendations.
- Describe the business process. "We need vendor bill approval for amounts over $10K" beats "help with AP."
- Include transaction volumes. "50 journal entries per month" vs. "5,000" changes the automation recommendation significantly.
- Ask about downstream impacts. Explicitly request how a change affects reporting, integrations, and other modules.
Limitations
The prompt does have some limitation, including:
- Knowledge cutoff. Recent NetSuite releases may not be covered. Verify against current documentation.
- No account access. Can\'t look up your field IDs, saved searches, or custom configuration.
- Navigation paths from memory. Paths are based on training data. Verify in your account, especially for heavily customized environments.
- No code generation. Provides functional specifications for developers, not production SuiteScript. (That\'s what the SuiteScript prompt is for.)
Bottom line: Treat this as a senior consultant who knows NetSuite finance cold but has never logged into your specific account. You still own sandbox testing and validation.
What\'s Next
The NetSuite virtual team is growing. The SuiteScript Developer handles code. The Finance Consultant handles configuration and accounting.
I\'m already working on the next one.
If you have a suggestion for which NetSuite role should be next, let me know in the comments.
This article was originally published on LinkedIn.