Most organisations have more possible AI projects than capacity. A portfolio approach helps choose well.
Discover Opportunities
- Interview teams about repetitive, text-heavy or prediction-heavy work.
- Review processes with high volume, delays or error rates.
- Look at customer pain points.
Assess Each Use Case
- Value: impact if successful.
- Feasibility: data availability, technical difficulty, integration.
- Risk: harm from errors, regulatory exposure.
- Readiness: a sponsoring team and clear owner.
Prioritise
Plot value against feasibility. Start with valuable, feasible, lower-risk projects to build capability and credibility.
Balance the Portfolio
Mix quick wins with longer-term strategic projects.
Stage Gates
Move projects from exploration to pilot to production with clear criteria at each stage, and stop projects that aren't working.
Reuse
Shared platforms, data and components make each new project cheaper.
Review Regularly
As AI capabilities change, previously infeasible use cases may become viable.