The situation
A global retail group had three distinct AI investment workstreams — and the board couldn't tell whether they reinforced each other or competed for the same dollars:
- An Asia-Pacific operational AI program, regional in scope.
- A Global Capability Centre buildout intended to centralize AI engineering.
- A BPO partnership with embedded AI deliverables, plus a joint venture structure with overlapping IP claims.
Each workstream had its own ROI case. None of them had CODN. The CFO was asking the question. The CEO needed an integrated view to defend the program at the next board cycle.
What I built
- Cross-region CODN model. Margin erosion + execution lag + talent flight + optionality decay, separately bounded for each region and partnership structure. Then re-aggregated at program level.
- Workstream interaction map. Where the three workstreams compounded (e.g., GCC capabilities accelerating APAC deployment), where they conflicted (e.g., JV IP claims complicating GCC IP ownership), and where they were redundant.
- Investment-defensible board memo. One document. Three workstreams. Bounded ROI and bounded CODN per leg, plus the integrated story.
- Governance instrument. Quarterly CODN drift review per workstream, with explicit triggers for workstream consolidation or expansion.
What changed
- Cross-region AI investment defended at board level as one program, not three competing asks.
- The JV IP-overlap question was forced into an explicit governance decision rather than ambiguity. Decision was clean and contractually bound.
- The GCC charter sharpened — it became the engineering arm for APAC's AI program rather than a parallel AI track.
- The BPO partnership added measurable AI-deliverable milestones with CODN-tied review gates, instead of relying on the partner's slideware.
What was hardest
Reconciling three separate finance teams' assumptions about depreciation, transfer pricing, and shared-services allocation across the JV. The CODN math is only defensible if the underlying finance assumptions are defensible — and they weren't until weeks of cross-team alignment work.
Stack used
Cross-workstream synthesis: Claude + structured CODN modeling templates. Shared with finance leadership directly; iterated through their assumption-correction passes before going to board.
Where this engagement shape fits
This was an executive advisory engagement — multi-month, board-deliverable, deep finance and legal interaction. Best fit when AI investment cuts across multiple geographies, partner structures, or P&L lines and needs a single defensible narrative. See engagement shapes →
Methodology
Run by the CODN framework. Engagements delivered through Stravonvale.