Read time: ~5 minutes · Series: Product Manager getting started · Next: vs Chat / BI
What this article helps you decide
Before rolling the tool out to the team, quickly judge whether Lantide solves your pain and who benefits first.
Good fit: teams and scenarios
1. Definitions get questioned repeatedly
Finance, ops, and business interpret the same KPI differently; weekly meetings ask "how was this number calculated?" If the answer lives in chat or someone's notebook, Lantide Plan + SQL + Report gives follow-ups a home.
2. Data is mainly local files or light connections
CSV / Excel / Parquet folders, or PostgreSQL / MySQL via ATTACH—no cloud warehouse required to start. Fits small teams, campaign retros, adhoc projects.
3. Want AI speed with human control
You accept Agent exploration and drafting but require Execute only by human, annotatable Plan, sign-off-ready delivery. That's governed Agent analysis, not a black box.
4. Small teams treat Plan as spec, Report as delivery
No dedicated data platform team; analysts pull data and tell the story; need one artifact for execution and reporting.
Poor fit scenarios
| Situation | Why |
|---|---|
| One-off charts, never questioned definitions | Generic chat BI may be lighter and faster |
| Expect in-app enterprise ETL scheduling | Lantide is a local analysis IDE, not a scheduler |
| Expect fully automated reports, no review | Opposite of Execute-by-human, Plan-must-review design |
| Team rejects SQL as definition contract | Product core is SQL-first |
| Occasional NL2SQL Q&A only | Overkill unless analysis will grow multi-step |
"Poor fit" doesn't mean never—first pilot should pick aligned scenarios for higher success.
By team size
| Size | Typical use |
|---|---|
| 1–3 analysts | Analyst owns Plan / Execute; PM occasionally annotates Plan |
| Analyst + PM + business | PM / business annotate Plan; analyst confirms before Execute; Report for meetings |
| With engineering support | Eng handles MCP / connections, .lantide migration—see Platform Admin intro; PM / analysts use Learn + USER_GUIDE onboarding |
Lantide doesn't require a dedicated "Agent admin"; governance comes from process (Execute, Queued Knowledge), not headcount.
Three signals a pilot is worth it
- In the last three months, at least one KPI dispute couldn't point to SQL or a document
- Analysts willing to write definitions in Plan (or willing to be pushed by process)
- Decision makers willing to spend 15 minutes reviewing Plan before Execute, not only "fast numbers"
Two or more true → schedule USER_GUIDE §2 demo or a real small project pilot.