Build Office REIT Lease Restructuring Simulator using AutoGen Multi-Agent Conversations
With commercial REITs facing impending debt defaults and going-concern risks, negotiating lease extensions and tenant retention terms is vital. Developers will build a multi-agent negotiation framework using AutoGen. The framework features a REIT Asset Manager Agent, a Tenant Representative Agent, and a Lender Agent. The agents engage in multi-turn automated dialogue to negotiate rent reductions, lease extension durations, and landlord tenant improvement (TI) allowances, seeking a consensus that avoids bankruptcy while preserving property cash flow.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an autonomous multi-agent negotiation simulator using AutoGen to model office lease restructurings under debt distress.
How work is evaluated
Evaluates negotiation efficiency, consensus reaching probability, and post-negotiation DSCR impact.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
DSCR Threshold Compliance
Verifies negotiated lease terms maintain REIT DSCR above minimum threshold (1.15)
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Consensus Rate
Rate of successfully reaching valid restructuring agreement • target: 0.9 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Configure AutoGen ConversationalAgents representing REIT Manager, Tenant, and Lender
Implement custom conversation turn limits, group chat management, and human-in-the-loop controls
Model cash flow optimization trade-offs between landlord concession costs and vacancy loss
Extract structured consensus lease agreements from multi-agent negotiation transcripts
Reference links and supporting material
Dataset of 30 commercial lease agreement scenarios experiencing lease expiration within 12 months under default risk.
Challenge setup
Evaluates negotiation efficiency, consensus reaching probability, and post-negotiation DSCR impact.
Challenge input
JSON with initial_rent_psf, remaining_lease_years, tenant_threat_level, reit_min_dscr
AutoGen
Required framework for multi-agent dynamic conversations.
Evaluated output
JSON with consensus_reached (boolean), new_rent_psf, extension_years, ti_allowance_psf, projected_dscr
- Verifies negotiated lease terms maintain REIT DSCR above minimum threshold (1.15)
- Rate of successfully reaching valid restructuring agreement • target: 0.9 • range: 0-1
- Benchmark: NegotiationBench 2025
- Consensus Rate target: 90%
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- AutoGen · Required
- Sarvam AI · Optional
- Confident AI · Optional
Evaluation contract
- DSCR Threshold Compliance · Weight 1
- Consensus Rate · Weight 1
Metrics
- Consensus Rate · Target 0.9
Benchmarks
- NegotiationBench 2025
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with initial_rent_psf, remaining_lease_years, tenant_threat_level, reit_min_dscr
Output: JSON with consensus_reached (boolean), new_rent_psf, extension_years, ti_allowance_psf, projected_dscr
Execution harness: Python
Evaluation tools
- AutoGen · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.