Vision
Vision
Fund Analyst Intelligence turns fund due diligence into a continuous, production-grade validation process.
It replaces periodic manual refresh cycles with monthly evidence-linked updates.
It makes every fund profile a living record, not a static document.
We are building the category of Continuous Fund Intelligence.
This category sits between research tools and governance systems.
It combines data pipelines, validation rules, and controlled narrative generation.
What we believe
1. Fund information must be treated as a changing system
Fund facts, risk characteristics, and operational details drift over time.
A fund profile is only trustworthy if it is continuously revalidated.
2. Evidence is a first-class product feature
Outputs must link to the underlying source material.
Every claim must be traceable to a document, a page, a timestamp, and a version.
3. Deterministic checks come before narrative
Numbers and hard fields must pass validation rules before any summary is produced.
The narrative must explain validated facts, not invent them.
4. Human approval is part of production, not a failure of automation
Regulated workflows require review and sign-off.
Automation should reduce work while keeping accountability.
5. Quality is operational, not rhetorical
A system is only valuable if it runs reliably every month.
The standard is repeatability, observability, and controlled failure modes.
The category we are building
Continuous Fund Intelligence
A platform that runs an end-to-end monthly cycle:
- Ingest user materials and trusted sources
- Extract and normalise fund fields and claims
- Validate against rules and historical baselines
- Detect material changes and exceptions
- Produce allocator-ready reports with provenance
This is not “AI research”.
This is a production pipeline for fund oversight.
Production goals
Reliability
The monthly cycle must run without manual babysitting.
Jobs must be recoverable and resumable by design.
Failures must be explicit, actionable, and logged.
Auditability
Every output must show what changed and why.
Every report must be reproducible from stored artefacts.
Every approval must be recorded with clear responsibility.
Consistency
The same inputs should produce the same validated facts.
Templates and rules should enforce stable structure and language.
Comparisons should be made against consistent historical baselines.
Signal over noise
The system must prioritise material changes.
It must suppress irrelevant churn and duplicate information.
It must produce concise exception lists that drive action.
Speed to insight
Monthly reporting should move from days to hours.
Analysts should spend time on judgement, not copy-paste.
What success looks like
A client can open any fund page and trust three things:
- The facts were validated in the current cycle
- The exceptions are material and explainable
- The report is allocator-ready and evidence-linked
That is the standard we are building towards.