| name | getting-started-with-research-analysis-in-strawberry |
| description | Help someone use Strawberry for open-ended research, market and ecosystem mapping, structured web data, synthesis, research artifacts, and recurring monitoring. Use when the user asks what Strawberry can do for research and analysis or has a broad research request without a clear starting workflow. |
Getting Started with Research and Analysis in Strawberry
Help the user see the larger research system, then begin with the piece of work that matters now.
Do not force every request through one methodology or make setup a prerequisite for a useful result.
Paint the research system
Show five connected kinds of work:
- Investigate a question: turn an open question into current, source-linked findings for a real
decision.
- Map a landscape: define the categories, entities, relationships, coverage, and patterns that
make a market or ecosystem legible.
- Build structured evidence: collect repeated fields from approved sources into a clean,
validated dataset.
- Explain the findings: shape the reviewed evidence into the brief, comparison, spreadsheet, or
deck the audience can use.
- Keep it current: preserve trusted sources and definitions, then monitor meaningful changes
once the method is stable.
These jobs can connect, but they are not a fixed sequence. Begin with the question, source set,
dataset, map, or artifact already in front of the user.
Choose a useful first result
Understand the decision, audience, freshness, scope, acceptable sources, exclusions, depth, and
useful output. Reuse relevant approved tabs, files, connected apps, prior research, and accepted
definitions before asking the user to reconstruct them.
If the starting point is unclear, suggest a small situational set, such as:
- investigate one decision-critical question and return a cited brief;
- map a market or ecosystem before choosing where to focus;
- extract a representative dataset from a defined set of websites; or
- research one company into a concise, source-linked brief.
Explain the breadth, depth, freshness, and cost tradeoffs that materially affect the work. Calibrate
with a varied first set when the taxonomy, schema, or research judgment is new and correction before
scale would save time or credits.
Route to the canonical owner
- Map a market or broad ecosystem: use
strawberry/research-analysis/map-a-market.
- Map a startup ecosystem: use
strawberry/venture-capital/map-startup-ecosystem when startups,
investors, and their relationships are the primary object.
- Compare marketing competitors: use
strawberry/marketing/run-competitor-analysis when the
specialist result is a competitor comparison for a marketing decision.
- Extract repeated web data: use
strawberry/research-analysis/extract-web-data.
- Research one company for a general decision: use
strawberry/research-analysis/research-a-company for a source-linked company brief without a
sales, marketing, or investment lens.
- Research a company for sales: use
strawberry/sales/research-an-account when the specialist
result is an account and relationship brief for a sales decision.
- Research an investment opportunity: use
strawberry/venture-capital/research-an-investment-opportunity when the specialist result is an
investment view.
- Turn accepted findings into a deck: use
strawberry/research-analysis/create-beautiful-slide-deck when the research already exists and the
remaining job is presentation production.
For open-ended research without a focused owner, complete the work normally. Keep source facts,
claims, transformations, inference, contradictions, uncertainty, and missing evidence distinct.
Make the next run better
Deliver the artifact at the depth and format the audience needs, with links and dates beside the
claims that matter. Preserve accepted source priorities, definitions, rubrics, schemas, formats,
and review points after the user trusts a real result.
Share the reviewed artifact when that is enough. Save a team method when colleagues should repeat
the same process. Offer a Routine only when the source set, trigger or cadence, change threshold,
audience, destination, review behavior, and stop conditions are stable.
