You are generating the AI Agent Regulator Overlap Mapping.

Inputs: SOC 2, ISO 27001, HIPAA control packs; EU AI Act compliance doc; NIST AI RMF mapping; DPIA; BAA/DPA addenda; regional regulator guidance (KE, NG, ZA, UG, RW).

Produce two files:

1. `AI_Agent_Regulator_Overlap_Mapping.md` — the narrative crosswalk and update triggers.
2. `AI_Agent_Regulator_Overlap_Matrix.md` — the wide matrix with one row per control area and one column per regime.

Hard rules:

- Cell content: the specific clause / criterion / standard ID for the regime, not a hand-wave.
- Where evidence is shared, name the artefact and the cadence.
- Where evidence is divergent (a regime requires something unique), name what is divergent and why.
- EU AI Act tier per feature explicit; prohibited / high-risk / limited-risk / minimal-risk.
- African DPA divergence explicit: KE residency expectations, NG NDPC advisory specifics, ZA POPIA s.71 explanation right, UG DPPA immediate breach standard, RW NCSA timeline.
- Update triggers: framework version change; new regulator advisory; new sub-processor in new jurisdiction.

Flag context gaps as `[CONTEXT-GAP: <area×regime> — <missing input>]`.