Sidekick

Analyzes natural language scheduling requests and automatically detects calendar conflicts by querying integrated calendar APIs (likely Google Calendar, Outlook). The system parses temporal expressions, participant availability, and timezone information to suggest optimal meeting slots without manual back-and-forth. Uses NLP to extract meeting duration, attendees, and preferences from conversational input rather than requiring structured form submission.

Generates draft email and message text based on user intent, then applies tone detection and style adjustments to match professional, casual, or empathetic registers. The system likely uses a fine-tuned language model to produce contextually appropriate business communication, with post-generation filtering to enforce tone consistency. Integrates with email clients or messaging platforms to surface suggestions inline or in a compose preview.

Provides a natural language interface to trigger scheduling, email composition, and other productivity tasks through chat commands. The chatbot uses intent classification to route user messages to appropriate backend services (calendar API, email generator, etc.), maintaining conversation context across multiple turns. Likely implements a state machine or slot-filling approach to handle multi-step workflows (e.g., 'schedule a meeting' → 'with whom?' → 'when?' → confirmation).

Analyzes participant calendars to identify free time windows and recommends optimal meeting slots based on constraints (duration, time-of-day preference, timezone). The system queries calendar APIs to fetch busy/free blocks, then applies heuristics or optimization algorithms to rank slots by suitability (e.g., avoiding back-to-back meetings, preferring morning slots). Results are presented as a ranked list of suggestions rather than requiring manual calendar inspection.

Generates complete email drafts from brief user descriptions of intent (e.g., 'ask John for a project update'). Uses a fine-tuned language model to produce contextually appropriate business email text, including greeting, body, and closing. The system infers formality level, recipient relationship, and email purpose from the input, then generates text that matches expected business communication norms.

Implements a freemium business model where core features (basic scheduling, email drafting, chat) are available free with usage limits, while advanced features (team collaboration, API access, advanced tone options) require paid subscription. The system tracks usage metrics (API calls, scheduling requests, draft generations) and surfaces upgrade prompts when users approach or exceed free tier limits. Likely uses feature flags to gate premium functionality.

Embeds Sidekick's chatbot and task automation capabilities into popular chat platforms via native integrations or webhooks. Users can invoke scheduling, email drafting, and other features directly from Slack/Teams/Discord without leaving their chat context. The integration likely uses slash commands (e.g., '/sidekick schedule') or @mentions to trigger Sidekick actions, with results posted back to the chat channel or as direct messages.

Classifies user messages into intent categories (scheduling, email drafting, general chat, etc.) to route requests to appropriate backend services. Uses a trained NLP model (likely transformer-based) to extract intent and entities (participants, dates, tone preferences) from conversational input. Handles ambiguous or multi-intent messages through clarification questions or fallback to general chat.