| name | source-investment-opportunities |
| description | Find and qualify a reviewable set of startups or private companies against an investor's thesis. Use when a venture investor wants new deal flow, a thesis-matched company list, lookalikes, emerging companies, or a repeatable sourcing workflow; keep market mapping, deep diligence, outreach, and CRM writes separate. |
Source Investment Opportunities
Turn the investor's real thesis into a sourced set of companies worth reviewing. The first result
should help the user correct the search, not bury them in an unexplained list.
1. Understand what belongs in the funnel
Clarify the sourcing goal and the decision the list should support. Learn the relevant stage,
geography, check size, sector, business model, ownership, traction, timing, and exclusions. Use the
firm's thesis, portfolio, passed opportunities, prior searches, CRM, and accepted examples when
available; do not ask the user to restate what those sources already establish.
Translate the thesis into observable signals and identify which criteria are hard requirements,
useful indicators, or questions that cannot be answered from public evidence. Do not turn a nuanced
thesis into a single opaque fit score.
2. Search beyond the obvious list
Use the sources that fit the search: current company sites, startup and funding databases available
in the logged-in browser, accelerator cohorts, investor portfolios, launch platforms, hiring,
founder writing, industry communities, research, news, and internal network or pipeline context.
Follow links through the visible browser so the user can inspect the path and take over when useful.
Search with plain language, adjacent categories, problem statements, technology changes, customer
types, and known lookalikes as well as standard industry labels. Preserve where each company came
from; source provenance can matter as much as a company field.
3. Verify, deduplicate, and qualify
Check that each company exists, is active enough for the agreed purpose, fits the stated scope, and
is not already in the portfolio, pipeline, pass history, or supplied list. Resolve likely duplicate
names and entities before presenting them.
For each candidate, gather only the evidence needed for preliminary review. This may include what
the company does, founding and location, stage and funding, business model, customer or traction
signals, team, recent changes, thesis fit, possible conflicts, and material unknowns. Label
inference and stale or weak evidence clearly.
4. Calibrate a varied first set
Present a manageable, varied sample before expanding a subjective or costly search. Include why
each company may fit, which evidence supports that view, and what remains unknown. Invite the user
to correct the thesis interpretation, sources, exclusions, fields, and ranking logic.
Expand only the accepted approach. A list is successful when it improves deal discovery and
judgment, not when it reaches an arbitrary row count.
5. Deliver and hand off cleanly
Return a reviewable list or table with source links, dates, fit rationale, uncertainties, and
deduplication status. Keep sourcing distinct from contacting a founder, requesting an introduction,
adding records, or moving pipeline stages.
Use strawberry/venture-capital/research-an-investment-opportunity to deepen a selected company.
Use strawberry/operations/make-a-warm-introduction when the next step is a credible network path.
Use strawberry/sales/keep-crm-updated for separately approved record changes.
After the user accepts the method, offer to save the thesis translation, sources, exclusions,
fields, and review behavior as a custom or team skill. A Routine may run on an agreed cadence,
deduplicate against the current pipeline, and prepare only new or materially changed candidates for
review. It should stop when the thesis changes, source access fails, or the search begins returning
mostly ambiguous matches.
