What is an AI augmentation score?

The score is a way to make the AI opportunity in work visible. Pulze Radar starts with an existing role and the processes behind it, breaking the work into tasks. The aim is to distinguish work that stays human-led, work that becomes hybrid and work where an agent can take the lead, with appropriate oversight.

An augmentation score is not a performance rating for an employee. It is not a probability of redundancy, a prediction of cash savings or a decision to remove a role. A role can contain tasks with very different opportunities. Reviewing the tasks gives leadership a clearer starting point than treating the whole job title as automatable.

Structure asks how clear the work is: the inputs, outputs and rules that guide a task. Repetition asks how often that work repeats. Tooling fit asks whether the tools, data access and system connections are ready. Risk surface asks what a mistake could cost and where human review and guardrails are needed.

These signals guide the assessment of work. Scores are capped at what the technology can reliably do today, rather than resting on promises about future capability. We do not publish a weighting formula on this website. The signals and score bands explain how to interpret the analysis, not how to calculate a score from an uploaded document yourself.

A score of 80 or above is AI-led. A score from 50 to 79 is hybrid. A score below 50 is human-led. These are interpretation bands on the 0 to 100 scale. AI-led does not mean no human involvement: ownership, review and decisions about the work remain part of the implementation.

Hybrid work combines AI support with the judgement and responsibilities of people. Human-led work keeps people at the centre. The distinction is about how a task is carried out, not the value of the person doing it. It helps leadership choose where to act and what the team will need to build and own the change.

The rings describe a readiness decision alongside the opportunity score. Adopt points to work and tooling ready for use with appropriate oversight. Trial calls for a bounded pilot before wider use. Assess means questions about the work, systems or risk need resolving. Hold means the constraints or risks must be addressed before proceeding.

A score and a ring answer different questions. The score shows the opportunity for AI support; the ring guides what to do next. A task with a high score can still require assessment or be held because a system cannot be reached or the right review is not in place. No automatic mapping between a numeric band and a ring is claimed here.

Yes. In the Modern Milkman case, 26 roles and 351 tasks were mapped. 97 tasks appeared automatable but were blocked by internal systems. Another 24 tasks were identified as build-first, with a roadmap of 10 business epics and 2 foundation epics across 3 waves.

That is why readiness belongs in Measure. Before building, we check the data foundation: its quality, access and connections to existing systems. The case shows identified opportunity and proposed delivery work. It does not publish realised client savings or establish that every task with an AI-led score is ready for production.

Pulze Radar is the engine of Measure in our method: Measure, Build, Embed and Workforce Transition. Its report includes task-level role maps, augmentation scores, tooling and integration readiness, risks and enablers, a roadmap in delivery waves and an ROI case per function. Leadership uses this to choose the first value case and investment priorities.

ROI uses the organisation's own roles, hours and costs. Our separate six-person finance model is illustrative only, not client data or a forecast. Build delivers working solutions with strategy consultants and AI engineers alongside your people. Embed establishes ownership and adoption. Workforce Transition puts better work, reskilling and redeployment first and supports a fair process where roles end, with HR and legal advisers.