How is this scored?

Every use case is placed on two axes. Return is the business value the reference organization can expect to realize, rated on the tiers below from Experimental up to Transformational. Investment is what it costs to get there across build, data and migration, governance, and adoption and change, rated from Already available up to Multi-year program. Each axis is split at a tier boundary to form the quadrants: Return of Meaningful or higher with Investment of Small team or lower lands in Invest, Return of Meaningful or higher with Investment of Dedicated team or higher in Plan, Return of Measurable or lower with Investment of Small team or lower in Experiment, and Return of Measurable or lower with Investment of Dedicated team or higher in Defer.

Return: business value delivered

TransformationalNew revenue stream, new product, or a materially different way of competing
Strategic20%+ across an entire operating function, or removes a structural constraint on growth
Major15–20% in a primary KPI: conversion, retention, throughput, cost to serve
Significant10–15% in a core KPI, or automates >70% of a named recurring process
Meaningful5–10% in a core KPI, or automates 40–70% of a named recurring process
Measurable2–5% in a core KPI, or automates a real but non-core process
IncrementalUnder 2% in a core KPI, or a clear improvement in a secondary KPI
OperationalReal time or quality gain, no clean line to a financial outcome
NarrowSmall share of volume or users, or sustained adoption unlikely
ExperimentalNo defensible business case today

Investment: implementation cost & complexity

Investment is the highest of the four factors, never the average.

Multi-year program
Build
Multi-year, new platform capability, permanent product team
Data & migration
New data infrastructure, ongoing ML-ops
Governance
Formal model risk management, regulatory review, audit trail
Adoption & change
Org design change, cross-department, exec sponsorship
Cross-functional program
Build
12–24 months, several teams across eng, data science, ops
Data & migration
New data platform components feeding several systems of record
Governance
Privacy, legal, security review; documented monitoring and failure modes
Adoption & change
New roles across more than one function, exec sponsorship
Multi-team project
Build
9–12 months, eng + data science + ops
Data & migration
New pipelines + integration with a system of record
Governance
Privacy and legal review, human in the loop, defined failure modes
Adoption & change
New roles or reworked workflows across a function, formal training
Extended team
Build
Squad + data/ML specialist, 6–9 months
Data & migration
New pipeline into one system of record, significant remediation
Governance
Privacy review with legal sign-off, human in the loop for exceptions
Adoption & change
Workflow change across several teams in one function, formal training
Dedicated team
Build
Dedicated squad, 3–6 months
Data & migration
Several new integrations, meaningful remediation
Governance
Standard privacy review, escalation and override paths
Adoption & change
Workflow change for one team, SOPs, manager reinforcement
Small team
Build
1–2 engineers, 1–3 months
Data & migration
Existing APIs, data clean enough
Governance
Covered by existing policy, low stakes
Adoption & change
Additive to current workflow, light training
Extended configuration
Build
Vendor config + light custom code, one engineer, 1–2 months
Data & migration
Minor mapping or cleanup within one system
Governance
Vendor assessment needed, low stakes
Adoption & change
Small change to one team's workflow, short training
Vendor configuration
Build
Moderate vendor config, 2–6 weeks
Data & migration
No new data work
Governance
Vendor already approved
Adoption & change
Self-evident to the people affected
Light integration
Build
Mostly configuration, days to a week
Data & migration
None
Governance
Normal change control
Adoption & change
None
Already available
Build
Native in the current platform
Data & migration
None
Governance
None
Adoption & change
Enablement and communication only

Multi-year program, Cross-functional program and Multi-team project are beyond the reference organization's internal capacity.

How the scores were produced. Each use case is evaluated with AI assistance against the tiers above: the model reasons through the business function, the two or three KPIs an organization would actually measure, the realistic uplift on those KPIs in a typical deployment, and what implementation requires in build, data, governance and adoption, then places the use case on each axis. Directional prioritization guidance for a reference mid-size commerce business. Not a financial projection or a commitment.