ProofThe Lighthouse Application goes live in real work.
Turns the opportunity chosen in the Diagnostic stage into a Lighthouse Application
The Proof stage turns the Lighthouse Application Decision and Plan from the Diagnostic stage into an application that works, is used, has its business value measured and is handed over to its business owner. That application is the Lighthouse Application.
The Diagnostic stage set the right starting point and the plan. The Proof stage shows, through the Lighthouse Application, whether that plan holds up in real life.
The Lighthouse Application is the first application that real users work with in real business conditions,and whose business value is shown under those conditions.
How the Proof stage runs
The Proof stage moves within the limits of the plan prepared in the Diagnostic stage; those limits exist to prevent complexity that would delay early value.
- 01
Week 1 · Clarification and design
Business problem, user scenario, scope, information sources, human oversight, security limits and success measures are confirmed one last time.
- 02
Week 2 · Building the Lighthouse Application
The solution, the information sources it needs and the core workflow are prepared; a first working version that users can try emerges.
- 03
Week 3 · Validation with users
Selected users work with the Lighthouse Application on real business cases; usability, output quality and workflow impact are validated.
- 04
Week 4 · Measuring value, handover and decision
Results are compared with the baseline. The Lighthouse Application is handed over to its business owner and moves into daily operation. Then the decision is made: move to the Embed stage, develop the application further, or stop. The handover does not depend on that decision; even if the organization stops, the application remains the organization's.
The Forward Deployed Engineer role →
How visible is AI literacy at this stage?
At this stage, AI literacy is not a separate training day; it develops with the application's real users.
In week one, a shared language of use and a safe starting frame are set. In weeks two and three, users build skills on real business cases: setting the right context, assessing output, spotting errors and knowing when to hand back to human judgment.
In week four, examples of correct use, approval points and good practices are recorded in the Lighthouse Application User Guide.
Outputs of the Proof stage
At the end of the Proof stage, the organization keeps:
- Lighthouse Application — an AI application that selected users work with in their real workflow, running within the agreed scope and handed over to its business owner.
- Usage and workflow design — at which step of the work, by whom and for what purpose the Lighthouse Application is used.
- AI literacy gained — users' experience of setting the right context, assessing output and using human oversight.
- Lighthouse Application User Guide — examples of correct use, security limits, approval points and good practices.
- Success and usage measurement — baseline data, usage findings, business value results and user feedback.
- Trust and governance learnings — conditions learned in real use on information, data, access, human oversight and compliance.
- Next-stage decision — an evidence-based basis for moving to the Embed stage, developing the application, or stopping.
What happens if the Proof stage succeeds?
Success is measured against the baseline and acceptance threshold set in the Diagnostic stage. Five areas are examined:
- Usage — are the target users working with the Lighthouse Application in their real work?
- Business value — is the targeted change happening in speed, quality, cost, revenue, customer experience, risk or decision quality?
- Output quality — are the AI outputs reliable, useful and good enough to support users' decisions?
- Trust and human oversight — do information security, compliance, authorization and human control points work correctly in real use?
- Ownership and learning — can the business owner track the Lighthouse Application, and can feedback and learnings carry into the next steps?
What the Proof stage does not do
- No demo, presentation prototype or technology trial is produced.
- Work does not go beyond the scope set in the Diagnostic stage.
- Software licenses, model usage fees and infrastructure investment are not included.
- Broad system integration and additional development the organization may request are not carried out.
Commercial terms of the Proof stage
The standard investment covers the design and build of the Lighthouse Application, user validation, AI literacy tied to the application, measurement, the user guide and the outcome decision.
Software licenses, model usage fees, infrastructure investment, broad system integration and additional development the organization may request are assessed separately.
Embed
The Proof stage shows that AI creates value on one selected business problem. The Embed stage turns that result into a repeatable way of working within the organization's people, workflows, decision mechanisms and management cadence.