Skip to main content

Uniphore Help Center Portal

2026.01 Self-Service Agent

Release Date: 09 February 2026

Note

Availability of this product is currently limited to selected customers. In addition, certain features described here may not be available for some customers. For more information, contact your Product Manager or Uniphore Sales representative.

The Self-Service Agent 2026.01 Release introduces the following features.

Introducing Uniphore's Agentic Self-Service Solution (Limited Availability)

Our new Self-Service Agent (SSA) is powered by a large language model capable of autonomous decision-making, with the ability to think, reason and act. It provides a human-like conversation experience for end-users, with natural-sounding contextual responses.

For complete details on creating and configuring the SSA, see the Self-Service Agent Guide.

Note

This next-generation SSA is now available in the DE production cluster.

Key Self-Service Agent Features

The SSA is based on the creation of intuitive AI Agents to support the following capabilities:

  • Personalized Opening Messages

    Each AI Agent supports complete personalization of the messages that open a conversation, leveraging end-user metadata that is available in telephony SIP headers or from pre-call API hooks. Personalized openings will result in higher containment and fewer unnecessary Agent handoffs.

  • Authenticate and Understand an End User

    Each AI Agent can use available tools to authenticate user identity (when needed) and understand user intent, generating responses based on the user's account-specific data as well as company knowledge base documentation.

  • Dynamic, Empathetic, Context-Aware Responses

    Each AI Agent can dynamically generate responses that are both empathetic and completely aware of the current context at every conversation turn. For example, the Agent might respond, "I understand your concern in not being able to make the payment last month…"

  • Guide an End User Towards a Business Goal

    Each AI Agent can guide the user toward a specific business outcome, handling all permutations of frequently-asked clarifying questions. For example, the Agent could suggest a product-oriented decision while focusing its answers on all of the related product benefits.

  • Guide an End User Away from Out-of-Scope Subjects

    Each AI Agent can firmly, but gently, guide the user away from conversations that are out-of-scope. For example, comparing data from a competitor, or attempts to perform an unrelated transaction such as paying a bill from another vendor.

  • Generate Responses Based on External Data Sources

    Each AI Agent can use available tools to access third-party sources (e.g., APIs and knowledge bases) and generate real-time responses based on the fetched data.

  • Gracefully Transfer a Call to a Human Agent

    Each AI Agent can monitor the conversation and, when required, intelligently transfer it to a human Agent together with its key metadata, complete context, and defined reasons for the transfer, e.g., failure to understand user input, failure to authenticate, etc.

Streamlined AI Agent Design

AI Agent designers can easily translate their ideas for creating a superior Customer Support experience into simple natural language instructions and component assignments for an AI Agent, and then let the SSA do the rest at run-time.

Using a straightforward, intuitive UI, designers quickly build and maintain AI Agents that define:

  • User persona and style of communication

  • Conversation flow

  • Session guardrails

  • Actions that the AI Agent should not perform

  • Questions that the AI Agent should not ask, and user questions that should not be answered

Designers can also create any number of third-party API integrations to be used as Tools to collect or submit key data for a session.

Each AI Agent supports uploading of knowledge bases or documents to the Uniphore Knowledge Layer to integrate with RAG capabilities, providing a powerful, scalable facility for question answering in real time.

Each AI Agent can be independently configured to define its voice response with ASR, TTS, no input and barge-in settings.

Pre- and Post-Call Hooks with Metadata

Each AI Agent is able to execute specific logic-driven actions, both at the start and the end of every call via assigned pre-call and post-call tools that use standard API calls. These hooks enable the SSA to proactively drive the session forward.

  • Pre-call APIs are useful for personalizing the greeting message and also to prime the session with specific user context. Pre-call APIs can provide key end-user metadata such as the possible reason for the call, a summary of previous interactions, the most recent transactions, and more.

  • Post-call APIs can be used to submit key dispositions, outcomes and call summaries to third-party systems for reporting purposes.

Constraints and Limitations

Item

Description

Workaround

1

A call intermittently terminates or transfers when the caller utters phrases such as “just hold on” or “please wait”. This issue is currently under investigation by the Development team.

None

2

The post-call hook sometimes fails to update the Conversation Summary after a chat session completes successfully, resulting in missing or outdated Summary data. This issue is currently under investigation by the Development team.

None

Bug Fixes

There are no defects specific to this Release.