Where inspiration meets reality
The executives we work with have studied their markets, tested assumptions with their teams, and formed clear views on where AI belongs in their business. They know what matters in their business better than anyone, and that is where this paper starts.
The thinking is rarely the constraint. The structure it has to travel through usually is.
The Earned Value Methodology is our answer to that pattern.
Three things set it apart
- 01
We scope by value chain, not by function or use case.
The unit of work is the outcome and every team it passes through.
- 02
The strategists stay through delivery.
The people who shape the work with leaders are still in the room when it ships.
- 03
We report what we retire.
Every use case is reported as realized, realized with feedback or retired, against a baseline agreed with your finance team.
Business outcomes do not sit inside one organization. They run through value chains: the processes, systems and teams that turn effort into revenue, margin and customer loyalty. Value runs across many teams, so the clearest view comes when leaders read the chain together. So a good idea has to cross many boundaries before it can move an outcome, and each crossing is a place where it can stall.
From idea to production
The ideas are already in the business. What turns them into outcomes is a fair test in the real business.
of AI proofs of concept the average organization scrapped before production, in S&P Global's 451 Research survey of 1,006 IT and business leaders.
We read those findings as a story about structure. Value is realized on the path between an idea and an outcome, and that path is where programs most often stall. Gartner published guidance in January 2026 that organizations which master the fundamentals turn pilots into production at twice the rate of those that do not.
Summarize a claim, draft a response, forecast demand, route a case.
Cost per claim, days to resolve, revenue per customer, losses avoided.
Workers in a large Danish study reported productivity benefits from generative AI, yet the authors ruled out effects on earnings and recorded hours larger than 2% two years after the launch of ChatGPT.
Use cases live in tasks. Outcomes live in the P&L. So the Earned Value Methodology starts from the outcome and works back. Every use case is tied to a business outcome and a value chain before anything is built.
Compute is rarely the binding constraint. Alignment and evidence across the value chain usually are.
AI takes the shape of the organization that builds it
Organized around the value chain, it delivers outcomes the whole business can see. Left alone, it will copy your org chart.
- Sales
- Operations
- Finance
- Risk
- Technology
- Partners
Left alone
AI shaped like the functions: sound pieces never designed to meet.
- Own data
- Own tools
- Own measures
Organized around the value chain
AI takes the shape of the value chain instead.
- Conway, 1968
Systems copy the org chart
Organizations which design systems are constrained to produce designs which copy their own communication structures. AI is no exception. Give each function its own AI program and you get AI shaped like the functions.
- The dynamo lesson
Redesign the floor
When factories first adopted electricity, many simply replaced the steam engine with a large electric motor. The real gains came later, when plants were redesigned around the new power source. Redesigning the work around the outcome is where the payoff lies.
- Orpheus Chamber Orchestra
Someone hears the whole
The orchestra plays without a conductor, and sends one or two players into the hall before each concert to hear the group as the audience will. Forward Deployed Strategists play that part for the value chain.
The Earned Value Methodology
It takes leaders' thinking to measurable business outcomes in four stages. It is judged by what it realizes and what it retires, and the people who shape the work are still in the room when it ships.
- Shared perspective
Leaders' own thinking brought together into one view.
- Connected value chains
Outcomes mapped across the teams that touch them.
- A cross-functional cohort
Value-chain owners, finance, risk, delivery teams and partners as one group.
- Trust by design
Controls and evidence built into the work from the start.
- 01
Speak value
AI Value AcceleratorOne week of light surveying, then a 2-hour session with executives and value-chain owners. Their time is scarce, and the format is built around that.
The gateValue brief
Target outcomes, baselines agreed with your finance team, a short list of experiments each with an owner, budget, time box and stop rule, and an agreed ownership and exit plan.
- 02
Launch and learn
Forward Deployed StrategistsA cross-functional cohort running prototypes built to test ROI, as measured experiments with real users, real data and real constraints.
The gateExperiment readout
Result against baseline for each idea, and the owner's decision to advance, change or retire.
- 03
Build what earned it
Forward Deployed EngineersProduction builds of only what proved itself in real use, with strategists staying in through delivery.
The gateProduction release
Evaluation results, a controls record and the first outcome readout.
- 04
Scale what drives impact
ScaleExtension across value chains through reusable patterns, knowledge transfer and handover.
The gateClient independence
Your teams run the next value chain without us. The honest readout decides what stays.
The honest readout
The honest readout shows executives what AI is worth to the business, in the outcomes they agreed.
of executives know where their AI revenue will come from, while nearly 80% expect AI to drive significant revenue by 2030, in IBM's 2026 CEO Study. The readout gives a leadership team that answer, one use case at a time.
Every use case is reported in one of three states, against a baseline and a threshold agreed in the Accelerator. Both the baseline and the readout are agreed with your finance team.
Budget, time box, stop rule and owner.
Against the baseline and the threshold. Agreed with your finance team.
Realized
It met or exceeded the agreed threshold against its baseline, and the readout is agreed with your finance team.
Realized with feedback
It moved the outcome measurably but has not yet reached the threshold. The cohort has named the change it will make and the date of the next readout.
Retired
It did not move the outcome within its time box, or its stop rule was triggered. The owner makes the call, and the work stops on purpose.
of experiments at Google and Bing generate positive results. That is the base rate, and it is why we test. Speed here means speed of learning, not speed of delivery.
Retiring work is only brave if it is cheap. That is why every experiment has a budget, a time box and a stop rule before it starts. A retired experiment costs what was agreed it could cost. A pilot that is never retired costs whatever it keeps costing.
Astro Teller, head of X, described giving bonuses to every person on teams that ended their own projects. Retiring work in the open deserves the same respect.
How many of the AI initiatives running in your business today could you retire with confidence?
Trust by design
Trust by design lets the business move with confidence. Controls are designed in from the start and evidence builds up as the work moves, so the people who answer for risk can say yes early. Risk and control owners sit in the Accelerator and help shape which ideas are worth testing. Experiments run within agreed guardrails, and production releases carry evaluation results and a controls record. In Scale, control designs are reused across value chains.
We've seen strong programs stopped late. In each case the teams were doing their jobs. The controls simply entered too late.
The first 90 days
Three moves help leadership teams take their thinking into the real business.
Three moves
- 01
Agree how the business speaks value.
Name the outcomes that matter most, how they are measured today, and the value chains they run through, with the leaders who own those chains and finance in the same room.
- 02
Get ideas into the wild.
Choose a short list of ideas tied to those outcomes and test them as measured experiments with the people who would use them. Agree in advance each one's budget, time box and stop rule, and who makes the call.
- 03
Build only what earned it, and report honestly.
Take the ideas that proved themselves to production with controls built in, and report every use case as realized, realized with feedback or retired.
References
- 01
S&P Global Market Intelligence, 451 Research
Voice of the Enterprise: AI and Machine Learning, Use Cases 2025. Online survey of 1,006 IT and business leaders, fielded 21 October to 25 November 2024. Reported in "Generative AI shows rapid growth but yields mixed results" (A. Johnston, 2025).
- 02
Humlum and Vestergaard
Humlum, A. and Vestergaard, E. "Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI." NBER Working Paper 33777.
- 03
Conway, 1968
Conway, M. E. "How Do Committees Invent?" Datamation, April 1968.
- 04
David, 1990
David, P. A. "The Dynamo and the Computer." American Economic Review 80(2), May 1990, pp. 355 to 361.
- 05
Seifter, 2001
Seifter, H. "The Conductor-less Orchestra." Leader to Leader 21, Summer 2001.
- 06
Kohavi and Thomke, 2017
Kohavi, R. and Thomke, S. "The Surprising Power of Online Experiments." Harvard Business Review, September to October 2017.
- 07
Teller, 2016
Teller, A. TED talk, 2016, as broadcast in "Astro Teller: When A Project Fails, Should The Workers Get A Bonus?" NPR, TED Radio Hour.
- 08
Gartner, January 2026
Chandrasekaran, A. "Why 50% of GenAI Projects Fail, and How to Beat the Odds." Gartner article, 26 January 2026.
- 09
IBM, 2026 CEO Study
IBM Institute for Business Value. 2026 CEO Study: Rewiring the C-suite, the fast track to 2030. May 2026. The revenue finding draws on IBM IBV, The enterprise in 2030, January 2026, a survey of 2,007 senior executives.
ibm.com/thought-leadership/institute-business-value/en-us/c-suite-study/ceo
CloudWerx is an AI-native consultancy that gives leaders' best thinking a way to be proven in the real business, and scales only what drives business outcomes.
To start a conversation, write to sales@cwx.tech.