Symphonia is a digital platform designed to improve how expert knowledge is synthesised for decision-making. It is based on the idea that evidence synthesis is a coordination problem: many experts hold partial information, but combining these inputs into a clear, usable output is costly and difficult.
In Symphonia, experts submit structured inputs (including positions, reasoning, evidence, and confidence). A large language model (LLM) then produces a draft synthesis that maps areas of agreement, disagreement, and uncertainty without making decisions or overriding contributors. This draft is reviewed and refined by participants through iterative feedback cycles, similar to a Delphi process.
The role of the LLM is not to replace expert judgement, but to reduce coordination bottlenecks by organising, compressing, and presenting large volumes of input in a transparent way. This allows larger and more diverse groups to contribute without increasing time or cost.
The output is a structured, versioned synthesis that preserves disagreement and supports evidence-based decision-making. This approach is particularly relevant for policy, healthcare, and research contexts where multiple perspectives must be integrated efficiently and rigorously.
