The right dimension pays twice
A small retention improvement compounds. A small reduction in remediation lag compounds. Value measured in the dimensions you own keeps paying long after the work ships.
The Virtual Engineering Value Framework
Measure the value, not the cost.
AI agents doing real software development change what an organisation can deliver. This is our framework for measuring that change: rigorously, qualitatively grounded, and in the language of the outcomes you are accountable for.
Classify · Map · Model · Tell
Multi-industry by design
01 / The strategic reframe
Ask what it multiplies, not what it costs.
Priced per task, virtual engineering is reduced to labour arbitrage, and every task is treated as interchangeable. That is the wrong frame. The right question is what happens to the outcomes your organisation is accountable for when its software ships faster, breaks less and stays secure.
The wrong question
- "What does virtual engineering cost per task?"
- Treats all engineering work as interchangeable
- Reduces the decision to labour arbitrage
- Invisible to the board
The right question
- "What happens to the outcomes we own?"
- Connects output to strategic priorities
- Positions engineering as a value multiplier
- Speaks the board's language
02 / The framework
Six value dimensions. You choose which matter.
Every unit of engineering output, and the pull request is the natural unit, can be classified by work type: features, fixes, security, refactoring, testing, documentation, infrastructure. Those categories roll up into broad value dimensions. Treat them as a menu you map to your own strategic priorities, whatever form they take: OKRs, mission outcomes, regulatory mandates or cost programmes.
Growth & mission outcomes
Capabilities ship sooner→
Faster to market, faster to serve→
Business won, citizens and patients served, mission advanced
Customer & user experience
Quality, reliability, performance→
Fewer failures, smoother journeys→
Satisfaction, loyalty, adoption
Risk, security & compliance
Continuous remediation & testing→
Audit-ready, always patched→
Avoided breaches, fines and loss of licence to operate
Speed & operational efficiency
Tech debt reduced, maintenance absorbed→
Capacity released→
Effort redirected to higher-value work
People & talent
Toil absorbed→
Engineers do the work they joined to do→
Lower attrition and replacement cost
Resilience & continuity
Knowledge captured, estates modernised→
No single point of human failure→
Continuity through change and departure
03 / The approach
Four steps, none of them generic.
-
01
Classify the output
Every pull request categorised by work type: features, fixes, security, refactoring, testing, documentation, infrastructure. The raw material of the analysis.
-
02
Map to your dimensions
Categories mapped to the value dimensions you actually care about, in whatever form your priorities take: OKRs, mission outcomes, regulatory mandates or cost programmes.
-
03
Model the mechanisms
Causal chains modelled using your own baseline numbers, never generic benchmarks. The value case is built from your data or it is not built at all.
-
04
Tell it in their language
Engineering leaders hear leverage, quality and tech debt. Executives and boards hear the outcomes they are measured on. Same evidence, two fluent tellings.
04 / Why it matters
Small improvements in the right dimension compound.
Standing still shows up too
Slower fixes, unpatched vulnerabilities and capability lag appear in exactly the same dimensions. The framework prices the do-nothing option honestly.
Measurement that directs the work
The same discipline that justifies the investment also steers it: point virtual engineering at the dimensions that matter most, and re-point it as priorities change.
Apply the framework to one month of your own engineering output.