Rapid prototype for agentic AI on ADB’s NSO workflow

ADB seeks innovative proposals for a rapid prototype that reimagines the Non sovereign Operations (NSO) deal lifecycle, covering deal origination, due diligence, contracting, and monitoring.

The Challenge

ADB seeks innovative proposals for a rapid prototype that reimagines the Non sovereign Operations (NSO) deal lifecycle, covering deal origination, due diligence, contracting, and monitoring. The prototype should demonstrate how deal information, documents, evidence, review activities, decisions, obligations, and AI-assisted outputs can be connected through a traceable and human-reviewed workflow. It must use synthetic data and is not intended to replace existing ADB systems or define a production solution.
 
This opportunity is under the ADB Digital Sandbox Program. Click here to learn more about available opportunities in the program.
 

Objectives

The objective is to prepare, configure, and demonstrate a lightweight working prototype for a prioritized NSO use case. The prototype should visualize a potential future-state workflow and enable ADB to assess the concept’s desirability, feasibility, and usefulness for further validation. It should also capture participant feedback and key observations from the demonstration.

 

EXPECTED DELIVERABLES

  • Prototype preparation and brief --- Reusable prototype components, a representative synthetic NSO deal scenario, and a concise brief defining the concept, primary persona, key workflow, representative inputs, and expected demonstration output.

  • Prototype walkthrough and feedback capture --- A demonstration using the representative synthetic scenario, including the capture of initial ADB feedback against the relevant validation questions.

 
 

  • Rapid Prototype --- A lightweight working prototype configured to the prioritized concept and agreed key user journey.

  • Close-out note and AI control matrix --- A concise summary of the prototype scope, feedback received, observed strengths and gaps, potential next steps, and a high-level AI control matrix for ADB’s internal reference.

 
Participate