Ada

Ada is an AI customer service agent platform for enterprises that resolves requests over chat, messaging, voice and email, priced per conversation.

Ada takes the routine requests out of a support queue, such as a refund, a cancellation or an address change, by putting an AI agent in front of customers on each channel a company runs. Ada Support Inc., a Toronto company founded in 2016, builds and sells it as its one product. The agent answers from a business's own knowledge, completes multi-step tasks and hands a conversation to a person when it needs one. Playbooks turn a written procedure, typed in plain language or uploaded as a PDF, into steps the agent follows. Knowledge comes from website imports that sync daily and from connected Zendesk and Salesforce knowledge bases. The Messaging product covers web and in-app chat, WhatsApp, SMS, Messenger, Instagram and custom channels. Ada Voice answers the phone with the same agent and escalates to a contact center. A persona setting shapes the agent's name, tone and style, and it detects a customer's language and replies in it. Handoffs pass the conversation and its context to live agents in Zendesk, Salesforce, Gorgias or Dixa. A homepage dashboard reports automated resolution, CSAT and cost savings, with topic and intent analysis underneath. The platform is designed for enterprise support teams that handle large volumes of repetitive contact. It suits operations whose requests follow several steps, where an agent that finishes the task saves more than one that only answers a question. Contact centers can add it to phone support, and companies with customers in several countries can serve each one in their own language. Buying starts with a demo, and the charge follows a conversation-based model in which each conversation the agent holds is billed, while some enterprises take a resolution-based option instead. A typical conversation opens in the website widget, the mobile app or WhatsApp, where a customer asks where an order is or wants to cancel. The agent works through the matching Playbook and checks the details through the systems it is connected to. Before a change reaches customers, teams run it through Interactive Testing and Simulations against the outcomes they expect. Developers pull conversation and message data out through the Data Export API, or drive conversations directly through the Conversations API. Administrators can also manage the agent from Claude through Ada's MCP server. When a case needs judgement, the human agent picks it up in the help desk the team already uses, with the full history attached.