Solutions / design disciplines

The disciplines I work across.

Eight disciplines I combine depending on what the outcome requires: four that design how the operation works—work, decisions, adoption and performance—and four that design the technology that runs it. A real problem rarely belongs to only one of them. They are lenses, not packaged offers.

01The disciplines, one at a timeOperating model · technology

01Process Architecture

Design the operation before designing the system.

I map how work actually flows—stages, owners, handoffs, decisions and exceptions—and design the target process before deciding which technology should run it.

  • Lifecycle design
  • Decision rights
  • Ownership and handoffs
  • Exception paths
Lead to customer, one lens at a time: Lead qualification, Potential customer, Opportunity, Commercial approval, Financial validation, Customer created, First order, Active customer.

Lead to customer, one lens at a time

  1. 01Lead qualification

    Owner: Marketing

    Decision

    Lead belongs to an existing customerRedirected to the account team

    CRM Lead scored & routed

  2. 02Potential customer

    Owner: Sales (handoff)

    Human action

    Probable duplicatePaused for match review

    CRM Account · prospectERP Party index lookup

  3. 03Opportunity

    Owner: Sales

    Human action

    Terms outside policyFlagged into approval

    CRM Opportunity & quote

  4. 04Commercial approval

    Owner: Commercial mgmt (handoff)

    Approval gate

    Rejected→ back to Opportunity

    CRM Approval & audit record

  5. 05Financial validation

    Owner: Finance (handoff)

    Approval gate

    Missing or invalid legal data→ back to Potential customer

    CRM Validation pendingERP Validation rules

  6. 06Customer created

    Owner: ERP (handoff)

    System action

    Timeout — outcome unknownHeld — reconcile, then continue

    CRM Number received Legal identity: CRM → ERP

  7. 07First order

    Owner: Sales (handoff)

    Human action

    Order exceeds credit limitHeld — credit review

    CRM Order status visibleERP Sales order

  8. 08Active customer

    Owner: Commercial mgmt (handoff)

    Lifecycle state

    Order cancelled before invoicing→ back to First order

    CRM Lifecycle · activeERP Invoice event

02Digital Operating Model

Transformation is broader than the technology it installs.

A commercial platform changes who decides, who owns which data and how performance is governed. I design those layers together, so the organization can run what it builds—and keep changing it.

  • Value streams
  • Decision rights
  • Governance and variation
  • KPI ownership
One change, followed through the operating model. The change “New rule: large discounts need finance approval” touches every layer: Process: A new approval step, with a time limit and a rejection path; Operating model: Finance gains a decision; sales loses one; Systems: An approval rule in the CRM and the quote tool; Data: Approval history becomes reportable evidence; Automation: Routing by threshold; the judgement stays human; Measurement: Approval ageing and discount leakage, each with an owner.

03Change & Adoption

Go-live is where adoption starts, not where it ends.

Systems change nothing until people work differently. I design adoption into the solution—roles, incentives, enablement and the signals that show whether the process is really followed.

  • Operating habits
  • Roles and incentives
  • Enablement on real work
  • Process compliance
Configured at go-live, adopted over time. Across Design, Enable, Adopt, Measure, Improve, the system is fully configured at go-live; “used as designed” rises gradually afterwards, while “left to training alone” levels off well below it.

04Systems & CRM Architecture

Give every system one clear job, owner and boundary.

I define system boundaries, responsibilities, sources of truth, identities, contracts and failure behavior—so CRM, ERP and the platforms around them each do one job well.

  • CRM architecture
  • ERP interaction
  • System ownership
  • Integration contracts
How a commercial process becomes a system architecture. CRM (Commercial) owns commercial lifecycle, approval record and the status users see; runs Lead qualification, Potential customer, Opportunity, Commercial approval. Integration layer (Platform) owns delivery state, idempotency keys, retries and correlation. ERP (Finance) owns legal and transactional customer, canonical customer number; runs Financial validation, Customer created, First order. All of them feed the data platform: analytical history and cross-system signals.

05Data & Integrations

Decide what moves, how fast, and who owns it.

I decide which system of record owns each fact, what moves by API, event or batch, and how both sides are reconciled when they disagree—before anyone maps a field.

  • Systems of record
  • API and event flows
  • Sync and reconciliation
  • Data quality
How commercial data moves between systems. CRM owns commercial lifecycle and pipeline; it feeds ERP through a synchronous link (API on request). ERP owns legal customer, orders and invoices; it feeds Data platform through a event-driven link (events as they happen). Data platform owns history and identity resolution; it feeds Analytics through a batch link (nightly batch). Analytics owns one published definition per metric. A daily reconciliation compares both sides of every flow.

06Automation

Automate the path, keep the judgement human.

I treat automation as process execution: which steps should run on their own, which need a person, and what happens when either of them fails.

  • Workflow orchestration
  • Approvals and escalation
  • Exception handling
  • Human-in-the-loop
A manual chain becomes an automated one—except the decision. Showing the automated version.
  1. TriggerManualAutomated

    Opportunity marked as won

  2. RuleManualAutomated

    Does this customer need a credit review?

  3. AutomationManualAutomated

    Creation request sent to the ERP

  4. Human decisionManualPerson decides

    Finance validates the credit terms

  5. Next stateManualAutomated

    Customer ready to order

The decision stays with a person. Everything around it moves on its own.

07AI & Agentic Workflows

AI that acts in real workflows, with people in control.

Not demos: agents that read structured context, take bounded actions in systems of record, pause for human approval where it matters, and leave an audit trail.

  • Structured context
  • Bounded system actions
  • Human approval
  • Audit and guardrails
An agent processes a request and stops at human review: Incoming email — RFQ from a distributor, three lines; Classify & extract — Reads it with permitted context only; Match entities — Customer matched; two of three products; CRM draft — Draft quote prepared—nothing committed; Human review — Approve, or send back with a reason. You are the reviewer.
  1. 01Incoming emailRFQ from a distributor, three lines
  2. 02Classify & extractReads it with permitted context only
  3. 03Match entitiesCustomer matched; two of three products
  4. 04CRM draftDraft quote prepared—nothing committed
  5. 05Human reviewApprove, or send back with a reason

Awaiting human review

08Revenue Operations

One commercial operating system, measured the same way everywhere.

RevOps connects process, CRM, data, targets and measurement into one system the business can steer by—with one definition per metric and a clear line from signal to action.

  • Pipeline and stage criteria
  • Forecast ownership
  • Management cadence
  • Performance governance
Five inputs, one commercial operating system: Process, CRM, Data, Targets, Measurement make up the commercial operating system. One KPI, Pipeline coverage: Definition — Open pipeline ÷ remaining target, by segment; Owner — Sales management; Computed — Data platform—published once, shown in the CRM; Triggered action — Add pipeline generation where coverage falls below plan.
02Discipline landscape8 disciplines · 2 groups · 6 layers

Discipline landscape

AOperating modelHow work, decisions, adoption and performance are designed
BTechnologyThe systems, data, automation and AI that run the operation
01 · Operating model

Process Architecture

Current-state analysis, target-state design and process governance: stages, decision points, ownership, handoffs, approvals, exceptions and SLAs—defined before anything is automated.

Operating model coverageProcessOperating modelSystemsDataAutomationMeasurement
  • Processprimary
  • Operating modelsupporting
  • Systemssupporting
  • Datasupporting
  • Automationsupporting
  • Measurementprimary
Shows up as
  • Approvals nobody really owns
  • Handoffs that lose context
  • Exceptions handled in email
Questions that change the design
  • Which decision does each approval actually make—and on what evidence?
  • Where does accountability transfer, and what must be true at that moment?
Explore Process Architecture
03Coverage matrixCapabilities × operating model layers
Capability01Process02Operating model03Systems04Data05Automation06Measurement
AOperating modelHow work, decisions, adoption and performance are designed
01Process ArchitectureCurrent-state analysis, target-state design and process governance: stages, decision points, ownership, handoffs, approvals, exceptions and SLAs—defined before anything is automated.PrimarySupportingSupportingSupportingSupportingPrimary
02Digital Operating ModelValue streams, ownership, decision rights, governance and measurement—designed together, then mapped to the processes, systems and data expected to support them.SupportingPrimarySupportingSupportingSupportingPrimary
03Change & AdoptionAdoption designed into the solution: operating habits, roles, incentives, enablement and the signals that show whether a new commercial process is actually followed.SupportingPrimarySupportingSupportingNot in scopePrimary
08Revenue OperationsPipeline stages, handoffs, forecasts and management cadence designed as one commercial operating system.PrimaryPrimarySupportingSupportingNot in scopePrimary
BTechnologyThe systems, data, automation and AI that run the operation
04Systems & CRM ArchitectureAn agreed customer lifecycle encoded in the CRM, and platform boundaries across CRM, ERP, integration and data—drawn from capabilities and ownership, not from the order in which tools were bought.SupportingSupportingPrimaryPrimarySupportingNot in scope
05Data & IntegrationsIdentity, contracts, state and recovery across CRM, ERP and the data platform—and operational data shaped into governed signals, each with an owner, a definition and an action it should change.SupportingSupportingPrimaryPrimarySupportingSupporting
06AutomationRouting, approvals, orchestration and lifecycle automation that remove coordination work without hiding policy, ownership or exceptions.SupportingSupportingSupportingNot in scopePrimaryNot in scope
07AI & Agentic WorkflowsAI workflows with explicit confidence, authority, observability and stopping points—people at the consequential boundary.SupportingNot in scopeSupportingSupportingPrimaryNot in scope
Primary—where the capability makes decisionsSupporting—shapes or is shaped by itOutside the capability
04Reading the columnsThe operating model stack
  1. 01Process

    Which stages, decisions and exceptions does the outcome require?

    Stage & decision model
  2. 02Operating model

    Who owns each stage, decision and handoff?

    Decision-rights matrix
  3. 03Systems

    Which system acts at each step—and which one is authoritative?

    Process-to-system map
  4. 04Data

    Which information is required, and who may create or change it?

    Data ownership matrix
  5. 05Automation

    Which work is deterministic enough to automate, and where must a person decide?

    Automation boundary
  6. 06Measurement

    How do we know the operation works—and who acts on the signal?

    KPI ownership model
05Method / Second LayerProcess · systems · data

The obvious requirement is only the first layer.

I make the assumptions underneath a request visible before they become expensive system behavior—whether the request is about a process, a system or data.

Apply the method to
Visible requirement
“We need an approval process.”
First layer

Add an approval step to the flow.

That is a step. It is not yet a decision model.

4 of 12 questions surfaced

Second layer — what the request leaves unsaid
  1. An approval without a named decision becomes a formality people learn to click through.