NetSuite work, rebuilt around AI.
A customization that took six weeks last year should take two by the end of this. Not because the work is easier — because the lifecycle is getting compressed at every stage.
Pricing the work the way 2015 priced it isn't going to hold.
The market has spent a decade pricing NetSuite work the way it priced NetSuite work in 2015 — three weeks of discovery, four weeks of build, a few days of sandbox clicking, and a thumbs-up. That pricing model is breaking.
Discovery, build, and testing are each getting compressed by AI. Not by 10%. By 50–70% each. And when each phase compresses, the next one starts sooner — which means the overall lifecycle doesn't speed up linearly. It compresses geometrically.
This is the methodology we're rebuilding our practice around. Here's what each phase looks like.
From scattered artifacts to a real spec
Traditional NetSuite discovery looks like this: three stakeholder meetings, a Slack thread, two PDF requirements docs, a folder of screenshots, and a spreadsheet of edge cases. Then a senior consultant spends a week trying to reconcile it all into a coherent specification.
With AI, all of that becomes a single context window.
Meeting transcripts, Slack exports, requirements docs, screenshots, sample data — fed into one place, with the right structure and prompting. The output:
- A first-draft technical specification in hours, not days
- Contradictions surfaced between what stakeholders said across different meetings
- Gaps and ambiguities flagged that the team didn't notice
- A document everyone can react to early, when changes are still cheap
The senior consultant still owns the spec. They're just editing a strong draft instead of writing from a blank page.
- Scattered docs, transcripts, screenshots
- Manual reconciliation
- A week of senior time before draft 1
- Single context window
- Draft spec in hours
- Contradictions surfaced early
SuiteScript generation with guardrails
Once the spec is locked, AI generates scaffolded SuiteScript 2.x — user events, scheduled scripts, RESTlets, Suitelets — directly from the requirements. What used to be days of boilerplate becomes hours of review and refinement.
The senior engineer still owns architecture, governance, and final accountability. The model writes the first 60–70%. The human writes the 30% that actually matters: the edge cases, the integration logic, the things AI doesn't know about your specific account.
It's not "AI codes for you." It's "AI handles the parts of coding that were never the hard part."
- Days of boilerplate from a blank page
- Repetitive Suitelet / UE / scheduled scaffolding
- Architecture and edge cases tangled together
- AI scaffolds 60–70%, human reviews
- Senior engineer owns architecture + governance
- Hours spent only on what's account-specific
The part everyone skips
This is the phase we're most excited about.
Most NetSuite customizations ship with manual sandbox testing — a few clicks, a thumbs-up, deploy. Real test coverage is rare because it's been too expensive to be worth doing.
AI changes that math entirely. Given a spec and the code, AI can generate:
- Test plans covering happy paths and realistic edge cases
- Synthetic test data sets across record types
- Automated validation that exercises the customization end to end
- Regression catches before anything reaches production
For the first time, comprehensive testing on small and mid-sized NetSuite projects is economically viable. Not because the technology is new — because the cost dropped enough that "good QA" can finally be the default, not the exception.
The next horizon: browser-driving AI agents that navigate a NetSuite sandbox the way a human tester would — open a sales order, fill the fields, save the record, verify the workflow fired, check the GL impact. The tooling is still early, but the trajectory is clear, and we're tracking the space closely.
- Manual sandbox clicks
- Thumbs-up, deploy, hope
- Real coverage too expensive to bother
- Test plans + synthetic data sets
- Automated end-to-end validation
- Regression catches before production
“When execution gets cheap, the cost of building the wrong thing gets cheap too. Bad decisions ship faster now.”
Judgment is what scales now.
AI commoditizes configuration and boilerplate. What gets more valuable is judgment — knowing what to build, where the gates and exception handlers need to live, when a stakeholder is asking for the wrong thing, which customization will become technical debt in eighteen months.
That's what senior NetSuite experience is for. And it's the part of the work that isn't getting cheaper.
The platform itself matters more in this world, not less. Its accounting integrity, its transaction model, the safeguards it brings out of the box — these become the foundation that prevents AI-generated work from quietly creating a mess. AI gets faster at scaffolding. The foundation underneath still has to be right.
What this means for what NetSuite work costs.
Each phase individually is 50–70% faster. That alone is significant. But the compounding is what changes the math.
When discovery compresses, build starts sooner. When build compresses, more time is left for real testing. When testing compresses, fewer bugs reach production — which means less rework, which means the next project starts sooner.
The customization that used to be quoted at $40,000 and six weeks is heading toward $15,000 and two weeks. The improvements that have been sitting on your “someday” backlog — too expensive, not strategic enough — are back on the table. The integrations you've been told would take a quarter? Closer to a few weeks.
We're heading there. Some engagements are already running at the new model. Others are in transition. The direction is set.
What still requires a human.
AI doesn't know which native NetSuite feature you're missing. It doesn't know that the workflow your stakeholder is asking for will conflict with the revenue recognition automation you put in last year. It doesn't know which “small customization” will become unmaintainable when you do your next acquisition. It doesn't know when to tell a client “you shouldn't build this at all.”
These are the calls that determine whether a customization is an asset or a liability eighteen months from now. They've always been the most valuable part of senior consulting work. AI makes them more valuable, not less — because more code is being shipped faster, by more people, with less oversight, and someone has to be the senior eye that catches what AI can't see.
That's the work we sell.
See the approach in practice.
The Document Extractor
A NetSuite customization we built that replaces a $12,000/year SaaS subscription. Built with this methodology. Shipped in two weeks.
See the productTalk to us about your project
Have a NetSuite customization in mind? Let's talk about what this methodology means for your scope and timeline.
Get in touch