Why voice changes everything
Reading a transcript of a debate is not the same as hearing it. Tone, hesitation, emphasis, interruption, these carry information that text alone cannot. When a synthetic persona pauses before answering a difficult question, that pause tells you something. When two personas talk over each other on a contentious point, the heat of that exchange tells you where the real tension lives.
Most AI simulation is text based. You read outputs on a screen. TownHall brings simulation to life through spoken dialogue. Synthetic personas engage in structured, spoken debate with distinct voices, natural cadence, and the emotional texture of real conversation. The difference is not cosmetic. It changes how decision makers engage with simulation outputs. People listen differently than they read.
What TownHall does
TownHall orchestrates multi agent deliberation through voice. You define a scenario, a policy question, an ethical dilemma, a strategic decision, and TownHall assembles a panel of synthetic personas with relevant expertise, diverse perspectives, and conflicting interests. Then they debate. Out loud. In real time.
Each debate is structured around established argumentation frameworks. Personas build on each other's arguments, challenge assumptions, introduce evidence, and negotiate positions. They do not just state opinions. They reason through disagreements the way real deliberative bodies do, with claims, warrants, rebuttals, and qualifications.
Live participation
TownHall is not just a spectator experience. Users can participate in debates through speech to text, asking questions, challenging positions, and introducing new information mid deliberation. The personas respond in real time, adapting their arguments to new inputs. Screen sharing with live vision analysis lets personas react to visual evidence, documents, data visualizations, and presentations shared during the session.
What you get
Every session produces a full transcript with speaker attribution, argument mapping, and key insight extraction. You see exactly where consensus formed, where it broke down, and why. The output is not a summary. It is a structured map of the deliberative landscape: which arguments were strongest, which positions were most contested, and which compromises might be possible.
Who uses it
Policy teams use TownHall to simulate public hearings before they happen, understanding which arguments will surface and which objections will be hardest to address. Ethics boards use it to stress test decisions across diverse moral frameworks. Executive teams use it to pressure test strategies with synthetic advisors who have no incentive to agree. Anyone making a decision that affects people with different values, interests, and priorities.
Key Capabilities
Structured Argumentation
Debates follow established argumentation theory. Claims are supported, challenged, and refined through genuine dialogue.
Distinct Voice Personas
Each persona gets a unique voice profile. You can hear who is speaking without looking at the screen.
Consensus Detection
Real time analysis of where agreement is forming, where it is breaking down, and the specific reasons why.
Full Session Recording
Complete transcripts, audio recordings, speaker attribution, argument maps, and extracted key insights.
Use Cases
Policy Deliberation
Simulate public hearings and stakeholder consultations. Hear the arguments before they happen in the real world.
Ethical Review
Stress test decisions across diverse moral frameworks. Surface concerns that internal teams might not raise.
Strategic War Gaming
Pressure test strategies with synthetic advisors who represent competitors, regulators, customers, and critics.
Training and Preparation
Prepare spokespeople and executives for difficult conversations by rehearsing against realistic opposing voices.
Interested in TownHall?
Join our Lab Partners program for early access and dedicated onboarding.
Become a Lab Partner