We utilise proprietary artificial intelligence to systematically identify worst-of autocall pairs where quantitative models overpay for historical volatility that no longer reflects actual forward-looking risk.
Structured product issuers (e.g., HSBC, Barclays, Citi, Morgan Stanley) price worst-of autocall coupons using automated engines driven primarily by trailing historical volatility. A stock that experienced a severe drawdown six months ago carries extreme trailing volatility — generating a rich coupon — even if the equity has since completely stabilised and the risk has passed.
This creates a systematic, repeatable spread between the coupon the bank is mathematically forced to pay (driven by past volatility) and the actual risk of barrier breach (visible in the forward-looking chart). Our AI visualises the chart and captures this spread.
Identify "blind spots" where trailing vol forces the bank to offer massive coupons, but forward-looking structural risk is actually low (e.g., a stock that crashed but has formed a firm base). The 50% barrier is set against a derisked floor.
Banks have completely different pricing for the exact same risk based on their internal derivatives books. By screening for $10B+ mega-caps and putting the structure out to competitive bidding among 5+ issuers, we naturally capture the top end of the spread.
When a bank's model flags extreme volatility and offers a >25% p.a. yield, the spread itself becomes capital protection. The accumulated income effectively absorbs potential downside risk at maturity.
Trailing 180-day volatility sees the crash. Pays a rich coupon automatically.
Stock has bottomed and formed a base. Barrier is far below the structural floor.
High coupon + low actual risk = systematic mispricing capture.
Investment banks rely on quantitative models. We use quantitative models plus advanced foundation models that visually analyse 5-year price charts — exactly as an experienced trader would, but applied simultaneously across 80,000+ pairs every week.
Before quantitative screening, creative models (Claude Sonnet) brainstorm non-obvious, cross-sector, and supply-chain relationships. This uncovers "hidden gem" pairs where yield is not yet priced out by crowded trades.
Every top-ranked pair undergoes a rigorous "Judge and Advocates" review. Fast, specialized AI agents (Claude Haiku) are strictly incentivised to build relentless Bear and Bull cases, eliminating AI bias and sycophancy.
A heavyweight reasoning model (Claude Opus or Amazon Nova) acts as the final Judge. It reviews the Bear/Bull debate alongside fundamental data and visually analyses 5-year and 1-year overlay charts to identify structural risks that raw volatility numbers miss entirely.
Each position is executed as a worst-of autocall on two US equities. The structure is strictly standardised across all positions to enable systematic, risk-managed portfolio construction.
| Parameter | Specification |
|---|---|
| Tenor | 6 or 12 months (Dynamic) |
| Knock-In Barrier | 50% European (Capital protected unless worst-of falls below 50% at maturity) |
| Autocall Trigger | 70% at quarterly observations (Early redemption if both stocks remain above 70%) |
| Coupon | Fixed, paid at maturity or autocall event |
| Underlying Assets | Worst-of on 2 US large/mid-cap equities |
| Minimum Ticket | EUR 500,000 per position |
| Reinvestment | Immediate upon autocall — driving continuous compounding |
| Equity Universe | 500+ US equities, $2B+ market cap, strictly screened for liquidity |
A single, consolidated strategy designed to exploit the pricing spread across the entire volatility spectrum, dynamically adjusting to market conditions.
A unified AI-driven strategy that exploits the spread between bank-priced volatility and forward-looking chart risk. The AI selects pairs across the full coupon spectrum — from conservative 8% anchors (stable names 10-30% off highs) to aggressive 20%+ yield plays (crashed names with high residual vol) — and builds a diversified portfolio targeting 10-14% gross yield with controlled breach risk.
The entire pipeline operates autonomously on a weekly cycle. Human oversight is strictly focused on model parameter governance and enforcing portfolio-level risk limits.
Creative models brainstorm thematic, cross-sector relationships to find non-obvious pairs. These are combined with a 500+ US equity universe and filtered for liquidity.
Pairs are scored mathematically on combined volatility, correlation, barrier breach probability, and bank-calibrated coupon estimates to find the best yield-to-risk ratios.
Fast, specialized AI agents act as relentless Bear and Bull advocates, generating targeted arguments for and against the pair's resilience to eliminate AI bias.
A heavyweight reasoning model (Opus/Nova) reviews the debate, fundamentals, and visually analyses 5-year price charts to make the final portfolio selection.
Rigorous AI-driven portfolio simulation from January 2016 to February 2026, utilising strictly point-in-time data. The simulation operates without hindsight bias — the AI is blind to calendar dates, relying exclusively on prevailing market statistics and price charts.
Key Achievements: Across a 10-year period encompassing the COVID-19 crash and the 2022 bear market, the strategy achieved a sustainable ~9-10% CAGR while maintaining an exceptionally low maximum drawdown of under 7%. By dynamically filtering for structural safety, the AI successfully deployed capital into hundreds of high-yield autocall positions with a near-zero barrier breach rate.
| Term | Detail |
|---|---|
| Vehicle | Managed Certificate or RAIF (Luxembourg) |
| Currency | EUR (USD share class available) |
| Minimum Investment | EUR 500,000 |
| Target AuM | EUR 10-50M initial |
| Management Fee | 1.0% p.a. |
| Performance Fee | 15% above 5% hurdle (High-water mark) |
| Liquidity | Monthly with 30-day notice |
| Counterparties | HSBC, Barclays, Citi, Natixis, Morgan Stanley |
The architecture is designed to scale across geographies: Global, US, EU, Asia ex-Japan, and Japan. The US strategy is fully operational and backtested. Other geographies follow the identical methodology with regional equity universes.
We are actively seeking strategic partners — investment banks, distributors, and anchor investors — to bring this institutional strategy to market.
Contact Management