Core Solution

The governed analytics platform for institutional investment teams

Built for the teams where Aladdin is out of reach and ChatGPT is banned by compliance. Alpha Quant Agent delivers governed workflows with deterministic execution, human approval checkpoints, and audit-ready output.

Business outcomes

Live platform — no implementation project required

The analytics catalog is in production. After a workflow diagnostic and onboarding session, your team runs its first governed workflow immediately — no multi-month integration timeline.

Standard reports produced every cycle, not rebuilt manually

Monthly factsheets, quarterly LP letters, GIPS composites, and risk committee packs — generated from a governed workflow, not assembled by hand.

Your proprietary models, governed and ready to act on

Register your house models — MATLAB, R, VBA, or Python — directly into the platform. They run with the same approval gates, anomaly detection, and audit trail as every built-in method. Your IP stays in your Azure tenant.

How it works

architecture

From portfolio question to governed output

Public overview of the workflow. Detailed runtime behavior remains available after login.

01

Bring in data and methods

Load the portfolio inputs and preparation baseline needed for the workflow.

02

Plan or approve the workflow

Choose fast execution or review-first planning based on stakeholder and governance needs.

03

Review outputs and artifacts

Receive analysis artifacts, method notes, and report-ready deliverables for stakeholders.

Execution modes

Direct execution

Move quickly when the request is clear and the workflow can proceed in one governed flow.

Plan-first execution

Review the proposed plan first, then approve execution when additional control is required.

What the agent can run

150+ governed analytics methods

Every method is deterministic, version-controlled, and runs through the same approval gate.

Sharpe / Sortino / CalmarBasic

Annualised risk-adjusted return ratios

CAGR & equity curveBasic

Compound growth and cumulative return chart

Drawdown analysisBasic

Depth, duration, and recovery per episode

Rolling volatility (20-day)Basic

Regime classification: low / medium / high vol

VaR & CVaR (historical)Basic

Tail-risk at configurable confidence level

Ulcer Index & QDDBasic

Investor journey and downside deviation metrics

GARCH / EWMA volatilityPro

Dynamic volatility model for risk overlays

Monte Carlo tail riskPro

Bootstrap + regime-switching simulated paths

House model integration

Bring your existing models into the governed workflow

Your team's proprietary models are often the most trusted signals you have. The problem isn't the model — it's that the output sits outside any governed process. Once registered, your model runs deterministically, its output is compared against your own run history, and every execution is permanently recorded.

  • Translate MATLAB, R, VBA, SAS, or Julia → Python with AI-assisted review
  • Every run compared against your own history — anomalies flagged automatically
  • Your code stays in your Azure tenant. Never shared with Nexqion or any third party.
See how house model integration works
MATLAB→ Python
R→ Python
VBA→ Python
SAS→ Python
Julia→ Python
PythonRegister directly

Client-registered models run in an isolated subprocess with a strict import whitelist. No network access, no filesystem writes.

Performance alpha — beyond reporting

The agent doesn't just report. It tells you when something is wrong.

After every run, three intelligence layers activate automatically — no dashboard to open, no threshold to remember to check.

Active after 5 runs

Cross-run anomaly detection

Every metric is benchmarked against your own run history. Anything beyond 2σ is flagged with context: '3.3σ below your 14-month average.' Activates after 5 runs.

Runs on every analysis

Mandate threshold alerting

Set beta caps, drawdown floors, VaR ceilings once in Settings. Every run checks automatically. Breaches are flagged, recorded in the audit trail, and surfaced in the Flags tab.

Every paragraph is cited

Compliance narrative

A plain-English commentary is generated from the computed numbers only — not from general knowledge. Every sentence cites the metric it came from. Defensible to a compliance officer.

What the output contains

One workflow. Seven analytics. One PDF.

Every governed run produces a structured report. Here is what a fund teaser pack output includes — the full report is shown on request.

  • 1Equity curve
    Chart
  • 2Sharpe / Sortino / Calmar
    Metrics
  • 3Drawdown table
    Top 5 episodes
  • 4Rolling volatility
    Regime chart
  • 5VaR & CVaR
    Tail risk
  • 6Factor attribution
    CAPM / FF3
  • 7Executive summary
    Plain-language memo
Sample output

Real reports. Download before the demo.

Generated from anonymised fund data. Open in browser or download — no signup required.

Animated preview

Inside the report pack

Auto-playing pages from the same PDFs linked below.

Fund Teaser
Fund TeaserPage 1 | Cover

Hover to pause | Click a dot to jump

LP outreach

Fund Teaser

1 page · A4 landscape

Single-page snapshot with 5 core KPIs, risk-adjusted verdict, and trailing returns. Sent to prospective LPs before the first meeting.

Open PDF
LP / Due diligence

Investment Pitchbook

9 pages · A4 portrait

11-year track record with peer universe comparison and investor suitability profile. Used in LP due diligence and initial allocation meetings.

Open PDF
Monthly reporting

Monthly Factsheet

8 pages · A4 portrait

Full performance attribution, rolling risk metrics, and fund-vs-benchmark analysis. Standard monthly reporting pack for institutional investors.

Open PDF
Quarterly letter

Investor Letter

7 pages · A4 portrait

Narrative-driven quarterly letter with market commentary, attribution analysis, and forward outlook. Designed for sophisticated LP communication.

Open PDF
Risk committee

Risk Oversight Report

8 pages · A4 portrait

COVID crash and rate shock stress scenarios with mandate compliance table. Used in monthly risk committee and board reporting.

Open PDF
Governance

Human approval gates — enforced at the API level

Before any sensitive analysis runs, Alpha Quant Agent pauses and requires an explicit human decision. This is not a UI suggestion — it is enforced at the API and worker level. Every approval is recorded with a timestamp and decision log.

01

Plan proposed

The agent proposes a structured JSON plan listing every method, parameter, and execution order.

02

Human approves or rejects

Execution cannot start without an explicit approval decision. Rejections are logged and recovery is prompted.

03

Execution + full audit trail

Methods run deterministically. Every step, decision, artifact, and timestamp is written to the audit record.

Plan approval required
Rejection logged with timestamp
Full audit trail per run
Enforced at API + worker level
UI-only suggestion
Override path exists

Regulatory bundles are force-gated to planned mode. The worker enforces this — no client-side override is possible.

Data security architecture

Your data, end to end

Azure Front Door + WAF

DDoS protection · OWASP rule sets · TLS 1.2+

01

Upload

CSV, Excel, or API feed

Encrypted in transit
02

Stored

Azure Blob Storage

AES-256 at rest
03

Processed

Isolated worker run

No persistence between runs
04

Output

PDF report + raw data

You own the output

Data region determined by your location

No data leaves your jurisdiction

Entra ID

Identity & access

Blob Storage

Client-isolated files

Container Apps

Isolated execution

PostgreSQL

Run logs & audit

Your data is not used to train models
No data shared across client tenants
No third-party tracking or analytics pixels