Voice, text, and visual interaction
The agent combines spoken interaction, written information, and visual content in the same conversation instead of forcing customers to use one medium throughout.
Multimodal conversations for customer service
Sierra’s multimodal agents are conversational AI agents that combine voice, text, and visual interfaces in a single customer interaction. Rather than keeping every step in one format, the agent adapts the interface to the task: customers can explain a need by voice, compare options visually, and use text when they need a reference.
The system is designed for customer-service workflows where conversation and structured interaction need to work together. A customer can discuss a disrupted flight, review alternate itineraries, select a seat from a map, or complete a form without restarting the conversation or repeating details. Teams build the agent once, add reusable visual components, and deploy the experience across the channels where the agent is available.
Core features
Context-aware mode selection
Sierra agents can select voice for explanations, visuals for side-by-side comparisons, or text for information customers may want to reference later.
Interactive product comparisons
Teams can present product cards and comparison tables so customers can review options together rather than relying on a representative to describe them one at a time.
Embedded interactive components
Calendars, seat maps, and forms can be embedded in the conversation, allowing customers to make selections or provide details directly in the interface.
Centralized component management
Teams design and host their own components, controlling how they look, what they show, and when they change. Updates are reflected wherever the component is used without separate platform versions.
Full-screen expansion
Larger interfaces can expand to full screen for calendars, long comparison tables, multi-step forms, and other content that needs additional room.
Practical use cases
Flight disruption and rebooking
An airline agent can show alternate flights with departure times, layovers, and prices after discussing a disrupted itinerary, then continue after the customer selects an option.
Seat selection
A customer can view a seat map during an airline conversation and tap the preferred seat instead of describing its location verbally.
Plan or product selection
A telecom or retail agent can discuss requirements by phone and show models, colors, storage sizes, and monthly rates for side-by-side comparison before the customer chooses.
Scheduling and form completion
A conversational workflow can use calendars for scheduling and multi-step forms when the customer needs to provide several pieces of information in a larger interface.
Pros and Cons
Pros
- Combines voice, text, and visuals within one continuous conversation.
- Lets customers compare options and make selections through interactive components rather than text or spoken descriptions alone.
- Supports reusable components across the agent’s deployed surfaces, reducing the need to maintain separate versions.
- Allows teams to control the design, content, and update behavior of embedded components.
- Can expand larger interfaces such as calendars and multi-step forms to full screen.
Cons
- The source does not provide pricing, plan limits, or detailed channel availability.
- Teams are responsible for designing and hosting the interactive components used in the experience.
- The source describes the concept and workflows but does not provide implementation-level technical specifications.
FAQ
What are Sierra’s multimodal agents?
Sierra’s multimodal agents combine voice, text, and visual components in one conversation. The agent can choose among these modes based on what the customer needs at each point in the interaction.
How are multimodal agents deployed?
A team builds the agent and can deploy it across the channels where the agent lives. Visual components can be reused across those surfaces rather than maintained as separate versions for each platform.
What interactive components can be added to a conversation?
Teams can use Sierra’s MCP UI integration to place interactive product cards, comparison tables, calendars, and forms directly in a conversation. The team designs and hosts these components and controls their appearance, content, and update behavior.
How does the experience handle changes between voice, text, and visuals?
Customers can switch between speaking, reading, viewing options, and tapping a selection without restarting the conversation or repeating previously provided information. Components that need more space can expand to full screen.
Quick Facts
- Product
- Sierra multimodal agents
- Category
- Conversational AI and customer service
- Interaction modes
- Voice, text, and visuals
- Interactive UI
- Product cards, comparison tables, calendars, forms, and seat maps
- Deployment model
- Build once and deploy across the agent’s available channels
- Provider
- Sierra
Sierra 替代品
Flunkey
Flunkey 是 Windows 的語音優先生產力工具，將口說內容轉成文字、AI 輔助輸出與可記住的上下文，讓你在不離開當前應用程式的情況下快速記錄想法與待辦。
PXZ AI
一個集成圖像、視頻、語音、寫作和聊天工具的全能AI平台，以增強創造力和協作。
Gemma AI
Gemma AI 是一款電話提醒 app，會依排程直接致電提醒你，不靠推播通知。支援 Google Calendar 同步與自然對話互動，讓你更直接掌握行程。
Wallie
Wallie 是開源 AI streamer，可觀看你的螢幕、聆聽聊天室，並以可設定的人設即時生成直播評論；支援本機執行與自有金鑰，適合無真人出鏡、自治直播與即時互動。
Claude Overlay
Claude Overlay 是適用於 Claude Code 的 Windows 桌面浮動疊加層，可讀取螢幕內容，讓你在不離開目前應用程式下提問、檢視內容並請求編輯。透過既有 Claude 訂閱與 Claude CLI 運作，無需額外 API 金鑰。
SpeakoFlow
SpeakoFlow 是適用於 Windows、macOS 與 Linux 的免費開源桌面語音轉文字應用，支援對任何 App 進行聽寫、Flow 語音寫作、AI 清理、翻譯與螢幕感知助理。