Our approach is to start with a demo and scope what’s best for your company
Our approach is to start with a demo and scope what’s best for your company
Our approach is to start with a demo and scope what’s best for your company
The data you provide stays in your environment. Any model we create is yours to keep; no obligation, no lock-in. The steps below show how a first engagement runs.
The data you provide stays in your environment. Any model we create is yours to keep; no obligation, no lock-in. The steps below show how a first engagement runs.
The data you provide stays in your environment. Any model we create is yours to keep; no obligation, no lock-in. The steps below show how a first engagement runs.
01
01
Entry Point
Entry Point
Entry Point
Initial video call where you bring any SOP, regulation or policy you care about.
Initial video call where you bring any SOP, regulation or policy you care about.
Initial video call where you bring any SOP, regulation or policy you care about.
02
02
First Meeting
First Meeting
First Meeting
Produce a live draft in seconds; discuss inputs, rules, outputs and confirm regulatory exposure.
Produce a live draft in seconds; discuss inputs, rules, outputs and confirm regulatory exposure.
Produce a live draft in seconds; discuss inputs, rules, outputs and confirm regulatory exposure.
03
03
Post Meeting
Post Meeting
Post Meeting
We will send a summary and draft workflow. If it makes sense, we can arrange a scoping workshop.
We will send a summary and draft workflow. If it makes sense, we can arrange a scoping workshop.
We will send a summary and draft workflow. If it makes sense, we can arrange a scoping workshop.
Initial Engagement
Initial Engagement
Initial Engagement
A workshop to scope your engagement and prepare a fixed-price statement of work.
A workshop to scope your engagement and prepare a fixed-price statement of work.
A workshop to scope your engagement and prepare a fixed-price statement of work.
Pilot Duration
Pilot Duration
Pilot Duration
A workshop and some review sessions over 1-2 weeks (depending on availability).
A workshop and some review sessions over 1-2 weeks (depending on availability).
A workshop and some review sessions over 1-2 weeks (depending on availability).
Included
Included
Included
Workshop minutes, any workflow models produced, and the SOW.
Workshop minutes, any workflow models produced, and the SOW.
Workshop minutes, any workflow models produced, and the SOW.
Not Included
Not Included
Not Included
Our code, a production model, a regulator walkthrough or handover. The diagnostic is to scope the work but not deliver it.
Our code, a production model, a regulator walkthrough or handover. The diagnostic is to scope the work but not deliver it.
Our code, a production model, a regulator walkthrough or handover. The diagnostic is to scope the work but not deliver it.
Data Storage Specifics
Data Storage Specifics
Data Storage Specifics
Layer
Layer
What it is
What it is
Where it lives
Where it lives
Standard
Standard
Domain Ontology
Domain Ontology
What it is
The vocabulary of the firm’s business: entities, relationships, rules
The vocabulary of the firm’s business: entities, relationships, rules
The vocabulary of the firm’s business: entities, relationships, rules
Where it lives
Inside the client’s environment, aligned to FIBO where applicable
Inside the client’s environment, aligned to FIBO where applicable
Inside the client’s environment, aligned to FIBO where applicable
Standard
OWL 2, RDF, FIBO
OWL 2, RDF, FIBO
Workflow Orchestration
Workflow Orchestration
What it is
The runtime that holds case state, routes work, and writes audit logs
The runtime that holds case state, routes work, and writes audit logs
The runtime that holds case state, routes work, and writes audit logs
Where it lives
Inside the client’s environment, client’s choice of engine
Inside the client’s environment, client’s choice of engine
Inside the client’s environment, client’s choice of engine
Standard
BPMN 2.0 (ISO 19510)
BPMN 2.0 (ISO 19510)
Decision Model
Decision Model
What it is
The executable rules: every input, every rule, every output
The executable rules: every input, every rule, every output
The executable rules: every input, every rule, every output
Where it lives
Inside the client’s environment, portable and vendor-neutral
Inside the client’s environment, portable and vendor-neutral
Inside the client’s environment, portable and vendor-neutral
Standard
DMN 1.5/1.6, FEEL
DMN 1.5/1.6, FEEL
Decision Engine
Decision Engine
What it is
The runtime that executes the model. Camunda, Drools, IBM ODM, OpenRules, or a custom in-house engine
The runtime that executes the model. Camunda, Drools, IBM ODM, OpenRules, or a custom in-house engine
The runtime that executes the model. Camunda, Drools, IBM ODM, OpenRules, or a custom in-house engine
Where it lives
Inside the client’s environment, client operates it
Inside the client’s environment, client operates it
Inside the client’s environment, client operates it
Standard
Open standard DMN runtime
Open standard DMN runtime
LLM Guardrail (Probabilistic Layer)
LLM Guardrail (Probabilistic Layer)
What it is
Reads, summarises, and classifies unstructured inputs before the decision model sees them. Does not make the decision
Reads, summarises, and classifies unstructured inputs before the decision model sees them. Does not make the decision
Reads, summarises, and classifies unstructured inputs before the decision model sees them. Does not make the decision
Where it lives
Inside the client’s private cloud tenant, against the client’s own LLM endpoint
Inside the client’s private cloud tenant, against the client’s own LLM endpoint
Inside the client’s private cloud tenant, against the client’s own LLM endpoint
Standard
Any LLM, client choice
Any LLM, client choice
Agent Integration
Agent Integration
What it is
How AI agents call firm systems
How AI agents call firm systems
How AI agents call firm systems
Where it lives
Inside the client’s environment
Inside the client’s environment
Inside the client’s environment
Standard
MCP (Model Context Protocol)
MCP (Model Context Protocol)
Provenance & Audit Trail
Provenance & Audit Trail
What it is
Every decision logged with input values, rules that fired, model version, timestamp, operator. Reproducible and inspectable
Every decision logged with input values, rules that fired, model version, timestamp, operator. Reproducible and inspectable
Every decision logged with input values, rules that fired, model version, timestamp, operator. Reproducible and inspectable
Where it lives
Inside the client’s environment, regulator access is whatever the client grants
Inside the client’s environment, regulator access is whatever the client grants
Inside the client’s environment, regulator access is whatever the client grants
Standard
DMN audit trail, BPMN event log
DMN audit trail, BPMN event log
Data
Data
What it is
Customer data, transaction data, decision logs, none of it leaves the client’s environment
Customer data, transaction data, decision logs, none of it leaves the client’s environment
Customer data, transaction data, decision logs, none of it leaves the client’s environment
Where it lives
Inside the client’s environment. Period.
Inside the client’s environment. Period.
Inside the client’s environment. Period.
Standard
—
—
Validation Corpus
Validation Corpus
What it is
The 12-month decision log the client supplies for back-testing
The 12-month decision log the client supplies for back-testing
The 12-month decision log the client supplies for back-testing
Where it lives
Inside the client’s environment, deleted after the engagement
Inside the client’s environment, deleted after the engagement
Inside the client’s environment, deleted after the engagement
Standard
Client-owned
Client-owned
Evidence Pack
Evidence Pack
What it is
The bundle for the regulator: model card, validation report, change log, governance documentation, plus an impact assessment where required
The bundle for the regulator: model card, validation report, change log, governance documentation, plus an impact assessment where required
The bundle for the regulator: model card, validation report, change log, governance documentation, plus an impact assessment where required
Where it lives
Delivered to the client, client owns it, submit directly to the regulator
Delivered to the client, client owns it, submit directly to the regulator
Delivered to the client, client owns it, submit directly to the regulator
Standard
Built to current regulatory literature.
Built to current regulatory literature.
Post Delivery Artifacts
Post Delivery Artifacts
Post Delivery Artifacts
Decision Model (DRD + FEEL)
Decision Model (DRD + FEEL)
Portable, vendor-neutral. Take it to any vendor, modify it, run it on any engine.
Portable, vendor-neutral. Take it to any vendor, modify it, run it on any engine.
Portable, vendor-neutral. Take it to any vendor, modify it, run it on any engine.
FEEL Tables
FEEL Tables
The rules, in plain-expression DMN notation. Edit them, audit them, re-version them.
The rules, in plain-expression DMN notation. Edit them, audit them, re-version them.
The rules, in plain-expression DMN notation. Edit them, audit them, re-version them.
Validation Report
Validation Report
The back-test against your 12-month decision log. Submit to the regulator, use as evidence.
The back-test against your 12-month decision log. Submit to the regulator, use as evidence.
The back-test against your 12-month decision log. Submit to the regulator, use as evidence.
Evidence Pack
Evidence Pack
The full model card, validation report, change log, and governance documentation. Submit to the regulator, hand to internal audit.
The full model card, validation report, change log, and governance documentation. Submit to the regulator, hand to internal audit.
The full model card, validation report, change log, and governance documentation. Submit to the regulator, hand to internal audit.
Delivery Models
Delivery Models
Delivery Models
The diagnostic produces a fixed-price SOW for the next step. These are the shapes we run most often. They are not a menu to buy from today.
The diagnostic produces a fixed-price SOW for the next step. These are the shapes we run most often. They are not a menu to buy from today.
Single-Model Pilot
Single-Model Pilot
2 to 3 weeks
2 to 3 weeks
2 to 3 weeks
One decision modelled, validated, and deployed. For when the diagnostic recommends one decision and you want to see a single production model before committing to more.
One decision modelled, validated, and deployed. For when the diagnostic recommends one decision and you want to see a single production model before committing to more.
One decision modelled, validated, and deployed. For when the diagnostic recommends one decision and you want to see a single production model before committing to more.
Multi-Model Rollout
Multi-Model Rollout
6 to 8 weeks
6 to 8 weeks
6 to 8 weeks
Three to five decisions modelled, validated, and deployed. For when the pilot is in production and you want to extend the same pattern to other decisions.
Three to five decisions modelled, validated, and deployed. For when the pilot is in production and you want to extend the same pattern to other decisions.
Three to five decisions modelled, validated, and deployed. For when the pilot is in production and you want to extend the same pattern to other decisions.
Sandbox Sprint
Sandbox Sprint
Schedule dependent on workshop output
Schedule dependent on workshop output
Schedule dependent on workshop output
Application, deployment, and graduation from the regulator’s AI sandbox programme. For when you want a regulator-validated use case in production before scaling.
Application, deployment, and graduation from the regulator’s AI sandbox programme. For when you want a regulator-validated use case in production before scaling.
Application, deployment, and graduation from the regulator’s AI sandbox programme. For when you want a regulator-validated use case in production before scaling.
Managed Retainer
Managed Retainer
Annual
Annual
Annual
Ongoing monitoring, change management, and quarterly validation. For when you want us to operate the model library, not just hand it over.
Ongoing monitoring, change management, and quarterly validation. For when you want us to operate the model library, not just hand it over.
Ongoing monitoring, change management, and quarterly validation. For when you want us to operate the model library, not just hand it over.
Implementation Examples
Implementation Examples
Implementation Examples
Four decisions we have modelled and validated against live decision logs. They are proof the methodology works, not a limit on what it can model. The standard is general-purpose: any decision with rules, inputs, and outputs.
Four decisions we have modelled and validated against live decision logs. They are proof the methodology works, not a limit on what it can model. The standard is general-purpose: any decision with rules, inputs, and outputs.
Credit Decisioning
Credit Decisioning
Approve, refer, or decline on retail and SME credit applications. Used by banks and lenders.
Approve, refer, or decline on retail and SME credit applications. Used by banks and lenders.
Approve, refer, or decline on retail and SME credit applications. Used by banks and lenders.
Pre-Trade Compliance
Pre-Trade Compliance
Block or release on pre-trade checks: suitability, concentration, mandate. Used by asset managers and SFC-regulated firms.
Block or release on pre-trade checks: suitability, concentration, mandate. Used by asset managers and SFC-regulated firms.
Block or release on pre-trade checks: suitability, concentration, mandate. Used by asset managers and SFC-regulated firms.
Family-Office Tax-Shield Monitoring (FIHV)
Family-Office Tax-Shield Monitoring (FIHV)
Eligibility check on the family investment holding vehicle concession (Schedule 16C plus the 2026 amendment). Used by single-family offices.
Eligibility check on the family investment holding vehicle concession (Schedule 16C plus the 2026 amendment). Used by single-family offices.
Eligibility check on the family investment holding vehicle concession (Schedule 16C plus the 2026 amendment). Used by single-family offices.
Alpha-Decay Detection
Alpha-Decay Detection
Triggers a review when a strategy’s realised alpha falls outside its historical envelope. Used by hedge funds and quant desks.
Triggers a review when a strategy’s realised alpha falls outside its historical envelope. Used by hedge funds and quant desks.
Triggers a review when a strategy’s realised alpha falls outside its historical envelope. Used by hedge funds and quant desks.