Delagent

Inspiration

From our work experience, we’ve seen firsthand how scheduling becomes a constant headache for executives. Juggling packed calendars, booking new meetings, and finding the right time for appointments, all while trying to maintain a healthy work-life balance, drains valuable time and energy. Many companies resort to hiring dedicated staff to handle this, but scaling that across multiple executives is expensive and inefficient. For organizations of any size, this quickly turns into a significant operational cost. In today’s technological world, there’s a smarter path forward: AI-powered scheduling agents. Instead of relying on human assistants for every executive, companies can deploy autonomous AI agents that negotiate meetings, protect focus time, and optimize schedules to deliver efficiency while dramatically reducing costs.

What it does

Delagent is a voice-first multi-agent scheduling platform. Each user has a personal AI agent that: Negotiates optimal meeting times with other agents. Explains its reasoning transparently. Protects individual focus and personal boundaries while finding group solutions. Integrates with voice commands so you can simply speak and your agent takes care of the rest. One-line Purpose: Personal AI agents negotiate meeting times between users, explain trade-offs in plain language, and protect individual focus time while finding optimal group solutions.

How we built it

Agents: Built autonomous scheduling agents using Fetch.AI’s uAgents framework, implementing negotiation protocols, conflict detection, and reasoning generation. Backend: FastAPI with PostgreSQL and Python. Frontend: NextJS and React interface showing live negotiations, calendar views, and voice transcription. Voice: Integrated OpenAI Whisper for speech-to-text, Text-to-Speech APIs for agent responses, and an intent extractor to parse meeting requests. Database: Fixed schema storing users, agents, calendars, meetings, and negotiation sessions to keep everything in sync across components.

Challenges we ran into

Integrating voice commands smoothly into the scheduling workflow was by far the msot difficult part of the project. We had to capture speech, transcribe it accurately, and then translate that into structured meeting requests all in real time.

Accomplishments that we're proud of

Achieved real-time updates where users can watch their agents negotiate live. Voice commands successfully convert into structured meeting requests, proving our vision for voice-first scheduling. Delivered all this within 24 hours, with agents, backend, frontend, and voice systems working together. Successfully building on Fetch.AI’s uAgents framework which allowed us to give each user their own autonomous scheduling agent that could negotiate on their behalf.

What we learned

How to use Fetch.AI uAgents for real multi-agent autonomy. The importance of clear protocols and data contracts when multiple components (agents, backend, frontend, voice) interact. How to balance user constraints (focus time, flexibility, preferences) with group optimization. The challenges of explainable AI, and why transparent reasoning is critical for user trust.

What's next for Delegant

Deeper voice integration, where your agent feels like a conversational partner you can talk to naturally, not just a command interface. Expansion beyond meetings into personal productivity orchestration: task management, reminders, travel planning, and coordination with other AI services. Scaling toward enterprise and cross-organization scheduling, where teams of agents negotiate on behalf of entire departments. Expanding from two agents into multiple with enhanced capabilities.