Intelligence

AI is the operating layer — not the decision-maker.

More signal. Less friction. Human accountability always.

Where AI Helps

Augmenting judgement. Not replacing it.

We use different frontier AI models for deeper company and industry research. AI expands coverage and speed — it does not replace human judgement or underwriting. Investment responsibility remains with people.

manage_search

Research Triage

AI synthesises earnings, filings, and market data across the coverage universe — surfacing what matters before the PM reads a single line.

monitor_heart

Portfolio Monitoring

Continuous position and risk monitoring with exception flags. Alerts the PM to drift, concentration, or news before it becomes a problem.

auto_mode

Workflow Automation

Routine tasks — reporting, data extraction, knowledge retrieval — are automated so the team spends time on decisions, not administration.

support_agent

Decision Support

Reviewable AI outputs feed into the investment process. PMs can inspect, interrogate, and override every recommendation.

Intelligent operating layerHuman sign-off required

database

DATAMarket & macro feeds

neurology

AI TRIAGESignal ranking & filtering

query_stats

RESEARCHContext synthesis

how_to_reg

DECISIONHuman accountability

Boundaries

What AI does not do.

AI does not make investment decisions. It does not set mandates, approve trades, or determine client suitability. Every decision that matters — buy, sell, size, risk — is made by a human with a name and a regulated licence.

Markets are not data problems. They are judgement problems. AI is extraordinarily good at the former. We retain humans for the latter.

person

Investment DecisionsHuman OnlyAll buy, sell, and sizing decisions are made by licensed portfolio managers — never by a model or algorithm

verified_user

AccountabilityNamed & LicensedEvery decision is traceable to a named individual who holds an SFC licence and can be held to account

preview

AI OutputsReviewablePortfolio managers can inspect, interrogate, and override every AI-generated recommendation at any time

assignment_ind

Client SuitabilityHuman AssessedSuitability assessments, mandate design, and risk profiling are conducted by the investment team — not inferred by AI

lock

Custody AccessNoneAI systems have no access to client accounts, funds, or custodian infrastructure — assets remain under independent control

edit_document

Mandate ChangesWritten OnlyClient mandates are set and amended by written agreement with the client — AI cannot alter investment scope or risk parameters

See the full investment process→

Human Accountability

Investment responsibility remains with people.

Ben Lin owns the High Beta and Quant books. Kenneth Li and Wyman Wong are named Responsible Officers under the SFC licence. The Investment Committee challenges every significant decision.

biotech

Research PhaseSynthesisAI aggregates data — PM reviews synthesised output and forms a view

how_to_reg

Decision PhaseHumanPM decides, sizes the position, and documents the rationale — AI does not participate

radar

Monitor PhaseContinuousAI monitors and flags exceptions — PM reviews every flag and decides whether to act

AI makes us better at our jobs. It does not do our jobs.

Intelligent infrastructure. Human accountability. Independent custody.

Begin a Conversation→