Liteo vs Referent

Liteo is a CRM for businesses that sell products or services and continue to support customers afterward. It keeps contacts, businesses, deals, customizable pipelines, and follow-up dates together. Teams can save products and services in a catalog, add them to deals, and send PDF quotes. For installed products, they can record the item, its location, service history, and next due date. Tasks, notes, reminders based on natural-language dates, website lead forms, and CSV import and export are also included. Integrations include Meta Lead Ads, Pabbly, and Zapier, with an API for custom connections. Liteo’s MCP server can connect to Claude, ChatGPT, Codex, or Cursor using the user’s Liteo permissions. It does not read or sync Gmail; optional Outlook sync is in beta and starts with new mail from the connection date. Liteo runs on the web and through an API. Annual pricing is $7 per user per month, billed at $84 per user a year, and includes a 14-day trial; monthly pricing is $9 per user per month. A lifetime seat costs $99 once.

Referent is AI-native practice management software for law firms, with agents handling work across intake, matters, email, deadlines, billing preparation, and follow-ups. Its intake workflow screens inquiries, opens matters, and prepares engagement materials for lawyer approval. After an inbox connection, it files emails, attachments, and context to the relevant matter. Users can direct work by voice. Lawyers retain approval control: client-facing and high-risk actions require explicit sign-off, and actions are recorded in an audit trail. Referent works with Gmail, Google Calendar, and Google Drive; Outlook and Microsoft 365 are marked as coming soon. An MCP server connects with compatible AI apps, while iPhone and Android apps let lawyers review agent work, approve actions, and check matters. Data is hosted on Google Cloud Platform with US and EU residency options, encryption at rest and in transit, encrypted backups, and per-firm workspace isolation. Referent says it does not train AI models on client or matter data. The service is in private beta or invitation-only access and is not an accounting system.