This is the scale we use to figure out where a company is with AI and what the next step actually is. It’s for one person’s proficiency — usually the owner or an executive who goes first. Collaboration between people comes later, once a few of them are at level 3 or 4. The edges between levels are fuzzy on purpose; real work doesn’t sort itself into neat rungs.
Level 1 — Learn how the tool works
Chatbot only. No connectors, no skills, no agents. The goal at this level isn’t output, it’s calibration: learning what these tools are good at, where they’re weak, and what it looks like when one is confidently wrong. Giving good instructions and context is a skill, and so is catching a bad answer. Most of the trouble people get into later comes from skipping this.
Level 2 — Connect it to NetSuite, read-only
Your own chatbot, connected to your own NetSuite account through the AI Connector Service, under a role that can see what you can see and nothing more. Ask for open sales orders, a customer’s balance, last month’s sales by rep, and get it back in seconds. Also at this level: uploading a spreadsheet, a contract or a report and working over it — often a non-technical exec’s first real “aha” — and saving a project setup so you stop re-explaining your company every time.
Level 3 — Simple automation
Scheduled tasks and basic tools: process the inbox, build a daily priorities list, pull the information you need out of a pile of documents. Tasks you learn to guide the AI through, for people who are juggling more responsibilities than hours. Skills — reusable instructions for a job you do repeatedly — fit here.
Level 4 — Use it to build something deterministic
A dashboard, a saved search, a script. The AI helps you build it, you test it, and once it’s verified there is no AI in the running of it. It does the same thing every time and you can audit it. This is where most of the durable value sits for a NetSuite company, and it’s the level people underestimate.
Level 5 — AI in the live information flow
The AI is inside the process itself, exercising judgment in real time: triaging incoming cases, routing them, drafting the response for a person to review. That’s the jump from AI-as-builder to AI-in-the-loop, and it’s where a customer-service assistant lives. Agents belong up here too. We’re skeptical that most executives get value from running coding agents; information-retrieval agents are more approachable and sit a little lower.
What we leave out
Fine-tuning models — too rare and too technical for this audience. Multi-agent orchestration — that’s a team topic, not a solo one, and belongs in the collaboration stage. And anything that depends on the AI drawing conclusions when what you really needed was for it to read a lot of data accurately. One of the best uses of AI in a NetSuite shop is seeing data no human could before — audit logs, log files, analytics — and that doesn’t need judgment at all, just scale.
Verification is a skill on every level
Learning to check AI output instead of trusting it is the habit that makes the rest safe. We teach it from level 1, and it’s what we do on every piece of AI work we deliver.
Not sure where you are? That’s usually a short conversation. Training and onboarding starts wherever you are.
The first hour is free. Bring us the thing that’s been bothering you and we’ll spend an hour on it with you, in your account, and send written notes afterward whether or not you hire us.