Research insights
Cursor AI and Lean Programming: Building Internal Tools Users Need

On this page
- What is Cursor AI and what is lean programming?
- Which internal tools can employees build with AI coding help?
- Why do internal tools need user research?
- What should you find out before and after building?
- How do you run a lean internal tool project step by step?
- What are the risks and common mistakes?
- FAQ
- Put users at the centre
Cursor is an AI-powered code editor that lets people describe what they want in plain language and get working code suggestions, which makes it possible for non-developers to build simple internal tools with much less coding. The tool is only half of the job: an internal app helps only if it solves a real problem for the people who will use it. A short round of user research before building, and a usage check after, keeps these lean projects useful and safe.
What is Cursor AI and what is lean programming?
Cursor is a code editor with built-in AI. Users can ask it to write, explain or change code, and it proposes edits across a project. It still produces real code, so it is “low-code with AI help” rather than truly no-code: someone has to review, test and run the result.
Lean programming applies lean thinking to software: build the smallest useful version, test it with real users, remove what is not used and improve in short cycles. AI coding tools fit this approach because they make the first version faster and cheaper to build.
Which internal tools can employees build with AI coding help?
- Simple forms and checklists that replace paper or long email chains.
- Small dashboards that pull data from spreadsheets or exports.
- Scripts that clean, merge or reformat files.
- Lookup tools, calculators and simple approval trackers.
Tools that handle sensitive personal data, payments or critical systems usually need IT involvement, security review and proper maintenance plans.
Why do internal tools need user research?
Internal tools fail for the same reason as customer products: they solve the wrong problem or do not fit how people work. Because AI makes building faster, it also becomes easier to build many tools that nobody uses. Asking users first keeps the effort focused on the steps that waste the most time.
What should you find out before and after building?
| Stage | Question | Simple method |
|---|---|---|
| Before | Which task wastes the most time or causes the most errors? | Short interviews and observation of the current process |
| Before | Who will use the tool, how often and on which device? | Interviews, short internal survey |
| Prototype | Can users complete the task without help? | Usability test with a few users |
| After launch | Is the tool used, and where do people get stuck? | Usage logs, short feedback survey |
| After launch | Did time, errors or rework go down? | Compare with the baseline measured before |
How do you run a lean internal tool project step by step?
- Define one problem. Write a one-page brief: the task, the users, what success looks like. The structure of a research brief works well here too.
- Talk to users. Run a few in-depth interviews and watch the current process.
- Build a small first version with AI help, covering only the core task.
- Test with real users and fix what confuses them.
- Review security and data handling with IT before wider use.
- Measure use and results, then improve or retire the tool.
What are the risks and common mistakes?
- Unreviewed code. AI suggestions can contain errors or insecure patterns. Someone must test the tool with real data.
- Data exposure. Check company rules before pasting internal or customer data into any AI tool.
- Tools without owners. A tool built by one employee can stop working when that person leaves. Document it and assign an owner.
- Skipping the “why”. Building first and asking users later leads to tools nobody adopts. Even internal surveys need neutral questions to get honest feedback.
FAQ
Is Cursor a no-code tool?
No. It is an AI-assisted code editor. It reduces how much code a person has to write by hand, but the output is still code that needs review and testing.
How many users should you interview before building an internal tool?
For a small internal tool, a handful of interviews with people who do the task every day usually shows the main problems. Add a short survey if the tool will serve many teams.
How do you know if an internal tool is working?
Measure a baseline before launch (time per task, errors, rework), then compare after launch alongside usage data and user feedback.
Put users at the centre
INTAGE Vietnam is a market research company in Vietnam and part of the INTAGE Group, Japan. The same principles we use in customer research, such as interviews, observation and well-designed surveys, help any team build tools people actually use. See our market research services or contact us.
Related reading
Latest posts
- Market Research: Building a Market Research Report for Investment Funds
Survey Sample Size: How Many Respondents Are Enough?
Survey Data Collection Methods in Vietnam: How to Choose- Technology Adoption in Enterprises: How to Research Employee Behavior
How to Write a Market Research Brief for an Accurate Quote


