The AI Agent in OmniServe is an AI-powered customer support agent that engages in direct message conversations with customers. It’s designed to answer questions, provide support, and deliver consistent responses using a brand-defined knowledge base.
AI Agent behavior in OmniServe
Where the AI Agent operates
Lucidya’s AI Agent operates within the OmniServe support flow by engaging directly in customer direct message (DM) conversations across supported messaging channels such as Facebook, Instagram, X, WhatsApp, and Live Chat. It processes customer messages in Arabic and English, accounting for Arabic dialect variation to maintain contextual clarity during conversations. Responses are delivered using a clear, service-oriented tone and follow consistent rules and structure across channels.
What guides its replies
The AI Agent’s responses are guided by a defined knowledge. This knowledge base is explicitly provided and may include written text files, approved URLs, and Swagger files that act as technical references and data endpoints. By operating within this provided knowledge, the AI Agent forms responses by referencing only verified sources, applying the same information consistently across customer interactions and channels.
The AI Agent’s behavior can be guided or customized to reflect a brand’s identity, influencing how and when responses are issued across one or more customer messages. This includes how responses are expressed in practice through tone, structure and phrasing, while the AI Agent operates within the conversation flow.
AI Agent knowledge base settings including uploaded URLs, text files, and Swagger files
What happens when the AI Agent hits its limits
When the AI Agent reaches the limit of what it can assist with such as when a request falls outside its configured knowledge or when a customer asks to speak with a human agent it initiates an escalation. This means the conversation is automatically handed over to a human agent for follow-up**.** A message is sent to let the customer know a human will take over, and from there, the engagement is routed based on the team’s assignment settings and the conversation is seamlessly handed over to a human agent for follow-up.
Where control is maintained
Control of the AI Agent is managed through clearly defined roles between agents and team leaders. While agents interact with the AI Agent as part of daily support operations, team leaders have extended access to settings and oversight capabilities that determine how the AI Agent is used within the support flow. This setup keeps AI usage controlled while keeping agent interactions simple. Key difference:
Team leader: Create, view, edit, pause, and resume the AI Agent. Full access to settings and inbox.
Agent: View the AI Agent’s settings and inbox. No access to create, edit, or manage the AI Agent.
Team leader and Agent access to AI Agent settings and permissions