The mechanism
Keep the report and the operating scenario distinct.
Read the submitted evidence, choose a bounded next step, document separate operating inputs, and review what actually changes.
01
Record the participant’s view
Each Diagnostic scores the structured answers within its instrument version. Keep the participant’s perspective and stated work context visible when interpreting that result.
02
Choose what to check
The report identifies recorded strengths and concerns and suggests bounded next steps where supported. It does not establish the cause of a reported difficulty.
03
Declare separate operating inputs
Use separately declared operational measurements and assumptions for financial planning. Value staff time as potential capacity, not as an automatic cash reduction.
04
Review actual outcomes
A later compatible Diagnostic can show changes in submitted answers and scores. Assess operational and financial changes using separate records; the Diagnostic alone cannot establish their cause.
Keep the evidence types separate: Diagnostic answers describe participant perspectives; operational records support activity measurements; financial assumptions define a scenario, not a proven return.
The comparison that matters
Three ways to ask "what is our bureaucracy costing us?"
Each has a legitimate use. A Diagnostic Instrument adds something the others do not automatically provide: a versioned baseline that can be measured again on compatible terms.
Participant responses describe reported experience. A separate operational scenario estimates potential capacity value and cash effects from declared inputs, not from the Diagnostic score. Read the Systems Measurement evidence.
A prompt
Ask a general AI chatbot
Genuinely useful for thinking out loud, exploring vocabulary, and pressure-testing intuitions. A capable model gives a good response. It is not an instrument.
- ✕A generated response is not, by itself, a reproducible measurement protocol.
- ✕A financial estimate still needs explicit operational inputs and assumptions.
- ✕Check which evidence supports each comparison.
- ✕Retain source records and review generated claims before using them.
A judgment
Commission a traditional advisory study
Experienced people, on site, rendering a considered judgment, with the implementation muscle and political navigation some conditions genuinely warrant.
- ~Agree the scope, method, deliverables and review process.
- ~Specify how later comparisons will use compatible evidence.
- ~Budget for the actual services, time and implementation support required.
- ✕Assess results using agreed operational records, not assumed financial return.
An instrument
Run the Monderman engine
A versioned diagnostic instrument: deterministic scoring, AI-assisted interpretation when enabled, and a record of the submitted answers for later compatible comparison.
- ✓Within the same instrument/scorer version, the same structured answers produce the same score.
- ✓Operational scenarios use separate activity measurements and declared assumptions; a Diagnostic score does not calculate savings.
- ✓Compatible comparisons can show differences in submitted answers and scores; they do not independently establish organizational change.
- ✓When an operational scenario is included, its measurements, assumptions, and costs remain visible for review.
To be direct about the boundary: some institutional conditions warrant a full advisory engagement: bespoke transformation, contested politics, sustained implementation support. Monderman does not replace that work. It can record participant perspectives before and after action where comparisons are compatible. Establishing operational change or financial return requires evidence beyond the Diagnostic.
Software supports repeated collection without commissioning a new advisory engagement each time. Choose the cadence and scope to suit the work, and keep the limits of participant-reported evidence visible.
Use a Diagnostic to organize reported findings and next steps. Use separate operational records and cost assumptions for financial planning; neither substitutes for an implementation study or a verified outcome.
The architecture
Why the read holds still.
Monderman splits the job in two, and keeps each half where it belongs.
A deterministic engine does the measuring.
Scores are computed by versioned, proprietary scoring logic from the structured answers. Within the same instrument/scorer version, the same structured answers produce the same score. Comparisons over time also require compatible scope, depth, and participant perspective; a changed score does not establish the cause.
That is what makes a baseline reproducible and a later compatible comparison meaningful.
AI assists at bounded points: never in score calculation.
When enabled, AI helps explain a completed Diagnostic or Synthesis result using selected evidence. Generated interpretation can contain errors and requires review. The current Diagnostic uses guided questions; interview mode is not available.
The model does not calculate or set the score. Within the same instrument/scorer version, identical structured answers therefore produce the same quantitative result even as language models change.
This division is what Monderman's research calls Deterministic AI Infrastructure (DAII): spend model capability only where judgment adds value, and hold everything that must be stable, measurement, comparison, cost, in deterministic systems. Scenario arithmetic uses the declared operational inputs and cost assumptions; language models do not set its financial factors. The full argument is published in From Tokens to Outcomes.