Meeting Users Where They Are
The traditional approach to B2B software involves forcing users to adopt new web applications or mobile apps. However, for certain utility workflows, the friction of 'app-switching' is too high. We identified scheduling as a prime candidate for conversational interfaces.
The Telegram Bot Architecture
We chose the Telegram ecosystem for its robust Bot API and highly engaged user base in crypto and tech communities. We built a bot that securely integrates with Google Calendar and Microsoft Outlook APIs.
Natural Language Understanding
Instead of rigid command structures (e.g., /schedule 2025-08-15 14:00), we integrated an LLM layer that understands natural language intents. A user can simply message the bot: "Set up a 30 min sync with the marketing team next Tuesday afternoon."
The bot then:
- Checks the calendars of all participants.
- Identifies overlapping free slots.
- Proposes times via inline Telegram buttons.
- Dispatches calendar invites automatically upon selection.
Performance and Scaling
To ensure sub-second response times, we utilized a serverless architecture (AWS Lambda) triggered by Telegram webhooks, backed by DynamoDB for fast user state management. The result is a seamless, frictionless scheduling experience embedded directly within the user's primary communication channel.