BLACKSMITH AI
AI-Native Equity Research Platform
INVESTOR BRIEF · April 2026
THE PROBLEM
Fundamental investing hasn't failed.
The process has hit a ceiling.
Today's analyst is expected to process thousands of filings, earnings transcripts, macro data, and alternative signals — across a coverage universe that compounds every quarter.
The result isn't better insight. It's inconsistency, fatigue, and structurally missed opportunities.
"The real inefficiency isn't mispricing — it's under-researched companies. Most value traps and overlooked opportunities appear clearly in filings. The constraint is who can process enough of them, consistently."
——————————————————
THE INSIGHT
The edge is shifting.
The advantage in fundamental investing is moving from who has the best individual insight — to who has the better research infrastructure. Firms that fail to upgrade will fall behind — not from poor judgment, but from limited processing bandwidth.
THE SOLUTION
A three-stage research funnel.
01 · QUANTITATIVE VALUATION ENGINE
Filters the entire market using deterministic, proprietary margin-of-safety models. No AI inference — fully reproducible.
02 · CONSTRAINED AI QUALITATIVE LAYER
Extracts evidence-based signals from company filings. Every output traces to a source document. No hallucination. No speculation.
03 · HUMAN JUDGMENT NODE
Final conviction decisions remain human-led. AI elevates the analyst above the bottleneck — it does not replace them.
Weeks of research compressed into hours — without sacrificing rigor.
——————————————————
EARLY VALIDATION · 2025
Live capital. Real compression.
~70%
Full-year portfolio return · 2025
<2hr
Research-to-conviction per company
Real capital deployed using the system — compressing full research workflows from days to hours. The key result was repeatable, decision-grade workflow compression.
Past performance is not indicative of future results. Generated in a personal account over a limited period; not independently audited.
CURRENT WORK
Regime validation in progress.
Stress-testing the system across historical market regimes — not to prove it works, but to precisely map where it performs strongest and how factor behavior shifts across conditions.
· High-rate environments
· Crisis & dislocation periods
· Liquidity-driven markets
· Value vs. growth cycles
A formal validation report will detail signal effectiveness, false positive rates, and regime-specific behavior — the critical step toward institutional-grade deployment.
——————————————————
WHY NOW
"Forward-looking investment firms are recognizing: the next edge is not better models — it's better research systems. As information volume compounds, research infrastructure becomes the durable competitive moat."
——————————————————
WHAT COMES NEXT
→ Completion of regime-based validation study
→ Expansion of proprietary factor framework
→ Multi-agent qualitative analysis refinement
→ Controlled onboarding of early strategic partners
Currently in a focused development phase — working with a limited number of partners while refining the architecture.
Want to see it in action?
I can walk you through how the system evaluates a real company — end to end — in a single session.
——————————————————
Blacksmith AI · Adam Thomas Contreras © 2026 All Rights Reserved
For informational purposes only. Not an offer to sell or solicitation of any security.
Do not distribute without authorization.