Multi-agent planner + evaluator + HITL — going operational in 2026.
Body
After 2024-2025, the year of demos, 2026 moves AI orchestration to production: specialised agents (geopolitical, macro, regulatory, adversarial, verifier) coordinated by a planner with a separate evaluator and human-in-the-loop gates on publishing. In olivLaw, every analysis goes through a chain of agents (collection, analysis, drafting, review, publishing), through deterministic gates that check figures, sources and forecasts, and through a human review before publishing.
Related
Knowledge graph reasoning
Model entities and relations to trace strategic dependencies.
ACH — Analysis of Competing Hypotheses
Reduces confirmation bias by simultaneously testing every plausible explanation.
Red teaming — adversarial challenge
Attack your own assumptions and forecasts from the adversary's perspective.