Staying power, not momentary change.
The real value of an AI engagement shows up three months later, not the day the training ends. The individual learns first; the workflow changes accordingly; the difference in business outcomes becomes measurable over time.
Three-layer measurement.
Three layers · L1 → L3- L3Where the value landsBusiness outcome
Revenue, cost, cycle time, customer experience.
- L2The process layerWorkflow output
Productivity, quality, throughput, error rate.
- L1The base layerIndividual capability
Maturity level, slope, AI Momentum component scores.
value flow: from capability to business outcome
Capability, output, outcome.
- L1Layer L1
Individual capability
The value of an AI program comes down to one question: who can do what at the end. Which maturity level the individual moved to within the AI Momentum frame, whether their slope is positive or a plateau, and which of the four components (mental model, strategy, build and production, responsibility) they lag behind in.
Sample metrics- Maturity level (L1–L4)
- Slope formula (T₂ − T₀) / Time
- Component scores (four axes)
Maturity slope · sample - L1
- L2
- L3
- L2Layer L2
Workflow output
The layer Workflow Rewire engagements measure directly. The productivity gain of an AI-embedded workflow, quality metrics, throughput increase, drop in error rate. Where the method touches the process, the measurement stays live.
Sample metrics- Cycle-time reduction
- Output quality score
- Automation rate
- Error percentage
Cycle time · sample 35−65% - L3Layer L3
Business outcome
What the line-of-business leader wants to see: the program's effect on the business. Revenue lift, cost reduction, compressed cycle time, improved customer experience. The layer where the engagement's ROI is validated.
Sample metrics- Revenue delta
- Cost delta
- NPS / CSAT change
- Decision-making time
Delta over three months · sample - Revenue+18%
- Cost−23%
- Cycle−41%
Measurement enters every engagement from the start.
T₀
During the engagement
T₂
The measurement instrument is put in place. Which metric, which instrument, which owner, which cadence.
Reviewed every week. The line showing how far we've moved from the baseline stays open at all times.
Slope or plateau. The engagement's real test; not the momentary effect at T₁.
Unfiltered feedback.
9.4–9.6
We run a standard feedback survey at the end of every engagement. The question set is structured across four dimensions (methodology clarity, content calibration, applicability, persistence); it also includes a likelihood-to-recommend question close to the NPS methodology.
The question set is identical in every engagement; that is what makes comparison between two engagements possible.
The most expensive step is starting in the wrong place.
In enterprise AI, most budgets evaporate on the wrong first step. The right starting point differs from company to company; a free 30-minute discovery call pins down yours.
