We build the data, signals, model, decision logic, agents and workflows, measurement, and handover system inside your environment. Each phase creates something useful. Each boundary gives you a clean choice to stop, hold, or continue.
Owning it sounds like the heavy lift. Built in phases, it is the lower-risk path: prove the first use case, keep the artefacts, then decide how far to go.
The build grows in stages, from the first use case out to a running operating state, and never leaves the boundary you own.
01 / OWNED LAYERS
Layers you keep. One capability your team runs.
Each is built in your environment and stays there. The stack below opens all seven layers, one at a time.
customer_data.layer
Data
Every customer signal cleaned, standardised, and governed in your environment. You hold the records, pipelines, and rules.
model_logic.layer
Model
The scoring, segmentation, prediction, prioritisation, or recommendation logic. Documented, versioned, and reproducible by your team.
decision_rules.layer
Decisions
Next-best actions, triggers, rules, and recommendations written as your logic and traceable back to the customer record.
workflows_measure.layer
Workflows and measurement
The lifecycle workflows, CRM motions, and frontline actions your team runs, with the AI agents working inside them. The measurement that shows what each move did.
capability_handover.runbook
Capability / Handover
Runbooks, documentation, training, operating rhythm, and internal confidence. The goal is not to make you depend on us. It is to make sure you could let us go.
Open any layer. See what it leaves behind.
OWNED LAYERS / DETAIL
Seven layers, built one on top of the next inside your own environment.
Open a layer to see what it does, the artefacts it leaves behind, and who
on your team runs it. When we leave, the whole stack stays with your team.
YOUR ENVIRONMENT7 layers · built in your house
STAYS WITH YOU
1Customer recordrecords.tbl
DOES
Unifies every customer into one governed profile you own.
MAKES
Records, identity keys, a data dictionary.
WHO
Data + CRM owners.
2Signalspipelines.yml
DOES
Turns first-party events into clean, named signals.
MAKES
Pipelines, event schemas, signal definitions.
WHO
Data engineering.
3Modelmodel.card
DOES
Scores propensity, value and risk from your own signals.
MAKES
Model cards, features, source history.
WHO
Analysts + data science.
4Decisionsrules.dsl
DOES
Encodes who gets what, when, as readable rules.
MAKES
Rules, thresholds, an audit trail.
WHO
Marketing ops + RevOps.
5Workflowsworkflows.map
DOES
Runs the journeys those decisions trigger, end to end.
MAKES
Workflow map, triggers, a message inventory.
WHO
Lifecycle + campaign teams.
6Measurementmetrics.def
DOES
Reports what each move did, against held-out controls.
MAKES
Metric definitions, dashboards, test logs.
WHO
Analytics + leadership.
7Handoverrunbook.md
DOES
Transfers the whole operating capability to your team.
MAKES
Runbook, documentation, training, access.
WHO
Your team, on their own.
Stays with your team
02 / THE FIRST 30 DAYS, THEN FIVE PHASES
Assess widely. Prove one thing.
Every engagement starts with a 30-day proof of concept: we map the decisions worth improving, from the customer record to the operating workflow, rank them, and build one POC against a bar you set before we start. The 30 days earn the build. If the evidence says go, the same method continues in phases, each with a clean stop point.
P1
Foundation
Question
What should your team own first?
Work
Data audit, commercial use-case selection, system map, first signal assessment, roadmap.
Across the five phases the question and the work are spent, while what you keep accumulates and stays yours.
A POC proves the idea on cases you chose, against a bar set before we built anything. It is not production, and it is not a promised return. It runs in your environment, or in a controlled NTWRK workspace where access requires it. The boundary is stated in the scope.
03 / THE OWNED LOOP
A loop, not a staircase.
Owned intelligence improves when each action creates fresh signal for the next decision. Because it runs next to your live data, the loop can act on what is true now, not only what was true last month. It compounds because your team holds the data and the decisions.
The return arc is the asset. Every cycle makes the next decision sharper, and the learning stays with you.
WHAT YOU END UP WITH
The method leaves you something to keep.
Records, logic, workflows, agents, documentation, and a team that can run them. Built in your environment, phase by phase, and yours when we leave.