Build a funded twelve-week enterprise AI portfolio roadmap.

All figures and evidence are fictional. A manufacturer has a $300,000 twelve-week investment budget and two engineers, each available for one engineer-week of work per week: 24 engineer-weeks total. Business owners and specialist reviewers are already assigned; their costs are included below. Work cannot exceed two engineer-weeks per calendar week. Each package includes its own implementation, evaluation, and handoff work. Budget and effort estimates are fixed for this exercise; assume no outsourcing or overtime.

A shared approved-data and integration foundation costs $40,000 and four engineer-weeks; it must finish before A or B starts. A is a demand-forecasting rollout costing $80,000 and eight engineer-weeks. Its completed pilot meets the agreed forecast-error and stockout thresholds in the intended rollout scope; finance supports $120,000 of projected annual net cash benefit, to be checked after rollout. B is a paid remote-monitoring service pilot costing $100,000 and ten engineer-weeks. Three design partners have signed paid-pilot commitments conditional on integration; a repeatable margin and broader demand remain unproven. C is a visual-inspection pilot costing $120,000 and twelve engineer-weeks. Labeled images exist, but performance under production lighting is untested; C does not need the shared foundation. Packages can be scheduled at one or two engineer-weeks per calendar week but must be completed in full to deliver their stated result.

The supply-chain director owns A, the product and sales directors jointly own B, the plant manager owns C, and IT owns the foundation. Management wants both operational value and evidence for a new revenue stream. No package is compulsory and unused funds may be retained.