Good sourcing starts with an operational mandate
An investment thesis document is not yet a sourcing specification. A living pipeline requires criteria expressed as observable filters, preferences, signals, and exclusions.
Europe is fragmented across languages, registries, programs, funding ecosystems, and disclosure practices. Provenance is therefore at least as important as reach.
1. Separate hard filters from soft preferences
Hard filters are criteria that make a company non-investable for the mandate, such as stage, ticket, geography, or a legal exclusion. Soft preferences change relevance without automatically removing the company.
- Sector and technology
- Stage, company age, and funding status
- Geography and legal presence
- Ticket, ownership, and reserve logic
- Team or business-model preferences
- Explicit exclusions and conflicts
2. Prioritize signals by decision value
Funding, formation, hiring, product, customer, grant, accelerator, research, event, and community signals can all matter. Their importance depends on the mandate and source reliability.
A signal should carry three facts: what happened, when, and according to which source. Only then should the pipeline explain why the event might matter to the mandate.
3. Treat source and freshness as part of the result
Sourcing results lose value quickly when source, timestamp, and conflicts are missing. Keep public, licensed, submitted, and permissioned information distinct.
OECD data shows strong regional and deal-size concentration in AI venture capital. Market reports like this provide context; they do not replace company-level provenance or a mandate-specific review.
4. Solve identity and deduplication before automation
One company can appear under multiple domains, spellings, former names, or program listings. The pipeline needs a stable external record ID and inspectable merge rules before save, pass, and evaluation decisions depend on it.
5. Run every pipeline as a learning loop
Track more than newly surfaced companies. Save, pass, evaluate, request information, and introduction decisions provide feedback about the mandate. One investment team's decisions remain private; aggregate patterns can improve retrieval without exposing notes.
A good pipeline gets smaller and more relevant as the mandate becomes clearer. More results are not success when analyst time per material opportunity increases.
Sources and context
The sources provide market and program context. The practical checklists are the FoundMatter working method and are not investment, legal, tax, or financial advice.