04 / Selected work
Strategy and applied AI: selected work from HEC Paris.
Two academic projects examine demographic strategy and the role of AI assistance in a complex transaction.
Selected academic work
Two projects, each built around a specific decision.
The demographic capstone tests how population change can affect long-term planning. HOMER asks where AI assistance could reduce friction in a real-estate transaction without removing professional review.
How I approached the work
Understand → Structure → Execute → Improve
Understand
Learn who is involved, what they need, and what limits the work.
Structure
Define the model, workflow, requirements, or plan.
Execute
Build the process, product, controls, and communication needed to deliver.
Improve
Test the result, use feedback, and fix what does not work.
01 / Demographic strategy
When demographic growth stops being the default.
The capstone asked how sustained fertility decline and population aging could alter strategic choices in two capital-intensive, population-sensitive sectors: real estate and resource extraction.
A changed planning baseline
From macro to sector
The same demographic shift creates different operating exposure.
Rather than treat population change as a single demand story, the analysis separated household formation, asset mix, labor availability, location, transition demand, and investment timing.
Sector exposure matrix
From evidence to action
Four responses for an uncertain planning environment.
- 01Rebalance portfolios
Test geographic and asset exposure against demographic scenarios.
- 02Adapt assets and demand models
Prioritize accessibility, retrofit logic, and changing customer needs.
- 03Build workforce resilience
Combine skills, inclusion, and fit-for-purpose automation.
- 04Institutionalize scenarios
Put demographic assumptions inside strategy and capital decisions.
02 / Academic concept
Where AI assistance could help in a complex real-estate transaction.
20-slide academic concept pitch
HOMER examined where controlled retrieval, drafting, and classification could reduce the customer’s workload. It also identified points where professional judgment, human review, and legal control would still be required.
Customer, AI-support, and human-control layers
- 01Search and preparation
- 02Plain-language guidance
- 03Visible next actions
Trust, comprehension, and exception handling
Retrieval quality, confidence, and escalation
Professional roles, partners, and market structure
Jurisdiction, responsibility, and compliant review