Demo
Phyn{Risk} · Enterprise risk management

Every risk number from one engine.

PhynRisk projects the cash flows of every loan, deposit, security and derivative once — from plain vanilla swaps to exotics — then calculates capital, earnings, liquidity, interest-rate risk, market risk and expected credit loss from them. The numbers agree because they are the same numbers.

See it runningWhat's inside ↓

Pipelines

Every task is a pipeline you compose.

Intake, checks, calculations and outputs are steps. Use them as they ship, write your own expression, or draw a decision flow, and place it in the pipeline.

Edit CRR3 SA-CR

Ready9 steps
Add a step

Your decision flow

Assign exposure class

CRR3-EXPCLASS-STD

Branching rules drawn on a canvas, then placed in the pipeline as one step.

CRR3-EXPCLASS-STDcanvas
bankcompanypersonnoyesStartCounterparty type?InstitutionSME?RetailCorporateCorporate SME

Path for a demo corporate exposure: 4 nodes visited, ending at Corporate.

Ready-made taskYour expressionYour decision flow
Illustrative · demo data

Runs you can see into

See every step of a run, its time, and where it stops.

Every task in PhynRisk is a pipeline of steps. When it runs, you watch each step, see how long it took, and when one fails, the error tells you which step and which data.

ALM Valuation — Meridia 2025-12-31 Banking book

Running
Pipeline
ALM Valuation
Analysis date
31-12-2025
Run
#829279
Started
11-10-2026 21:57:04
Duration
0.0s
Type
Process Pipeline

0 of 7 steps

  1. Data

    Load contractual data

    —

  2. Data

    Load market inputs

    —

  3. Data

    Set up market

    —

  4. Foundation

    Build payment schedule

    —

  5. Foundation

    Generate cash flows

    —

  6. Foundation

    Compute NPV

    —

  7. Foundation

    Persist results

    —

01

Every step, in order

The run map shows each step of the pipeline with its status, as it runs.

02

Where the time goes

Time per step and a waterfall view: a bottleneck is visible, not guessed.

03

Errors point to the place

A failure names the step, the box or the line, and the data behind it.

Illustrative durations · demo data

Pipelines · agent-ready

Give it a task, not a question.

Pipelines and rules are configuration, so an agent can read one, change it, test it and show the impact. On PhynRisk it does the work, and you sign off.

Pipeline · CRR3 SA-CR · Live pack

Reading

8 steps, run left to right. The agent opens the step that sets the corporate risk weight.

  1. Initialise context
  2. Load contractual data
  3. Load collateral data
  4. Credit risk mitigation
  5. Assign exposure class
  6. Assign risk weight CRR3-RW-CORP
  7. Generate exposure results
  8. Summary results

CRR3 example on the live pack · illustrative · demo data

ONE ENGINE

Calculated once. Used everywhere.

Your data

Positions, customers and market data, read from your existing systems, not replaced.

The engine

Contractual and behavioural cash flows, present values, curves and scenarios.

The measures

Capital, NII, gap and liquidity, IRRBB and FTP, IFRS 9, market risk.

Your processes

Regulatory returns, ICAAP/ILAAP, LCR and NSFR, stress testing, hedge accounting.

Feed your existing reporting platform, or produce the returns here.

WHAT'S INSIDE

Everything a risk function has to produce.

Cash-flow and pricing engine

Contractual and behavioural cash flows and present values, instrument by instrument.

Derivatives

Plain vanilla and exotic derivatives, priced on the same engine for NPV, sensitivities and risk.

Regulatory capital

Risk-weighted assets and the capital requirement under CRR3 and Basel 3.1.

NII and earnings

Forward net interest income and multi-year P&L.

Gap and liquidity

Contractual and behavioural gaps, bucket by bucket.

IRRBB and FTP

ΔEVE, ΔNII, supervisory outlier tests, transfer pricing and margin attribution.

IFRS 9 ECL

Staging and expected credit loss on the same cash flows.

Market risk and FRTB

Sensitivities, VaR and the standardised approach.

Dynamic budgeting

Balance-sheet plans projected forward.

Reporting and queries

Regulatory returns, extracts and ad-hoc questions from the same results.

Built on shared behavioural models, market and funding curves and a scenario generator.

A NEW WAY OF WORKING

Describe the change. The agent builds it, you sign it off.

In PhynRisk, regulation is a decision graph, not rules buried in compiled code — the CRR3 standardised approach, for example, with every node linked to the article it implements. Describe a change in plain terms; an agent builds it on the graph, citing the article it relies on, and tests it in seconds. You review and sign it off before it applies — and a supervisor's question leads to a path you can point at.

  • Decision nodes, each linked to its article in regulation
  • Article-level test cases run in seconds
  • Every change built by the agent, signed off by a person, and recorded
CRR3
Agent identifies the decision logic
ExposureArt 114Art 119Art 121RWRWRW
Tests passedAwaiting sign-off
UNDER YOUR CONTROL

AI does the work. Your people make the calls.

Numbers come from the engine, not the model.

Calculations are deterministic: the same inputs always give the same answer.

The agent drafts, a person approves.

Nothing reaches production until someone accepts it, and that acceptance is recorded.

Every figure traces back.

Any number resolves to the positions, curves and parameters behind it, and can be rebuilt for a supervisor years later.

See PhynRisk on your own numbers.

Design partners get PhynRisk configured on their own portfolio, working directly with the people who built it.

See it running

45 minutes with the people who built it. No slides.