Merge pull request #3102 from dineshr1493/cms/dineshr1493/hpe-dev-portal/blog/part-6-agentic-ai-teams-in-router-mode-multilingual-routing-with-agno

Create Blog “part-6-agentic-ai-teams-in-router-mode-multilingual-routing-with-agno”

Author: choukden

content/blog/part-6-agentic-ai-teams-in-router-mode-multilingual-routing-with-agno.md

@@ -0,0 +1,203 @@
+---
+title: "Part 6: Agentic AI teams in Router Mode: Multilingual routing with AGNO"
+date: 2025-07-21T10:04:38.579Z
+author: Dinesh R Singh
+authorimage: /img/dinesh-192-192.jpg
+disable: false
+tags:
+  - "Language Agents "
+  - MultiLingual Application
+  - Multilingual
+  - Multi Language Agents
+---
+<style>
+li {
+   font-size: 27px;
+   line-height: 33px;
+   max-width: none;
+}
+</style>
+
+One of the most powerful capabilities in Agentic AI is orchestrating multiple agents to work together—each with its own specialization. In this segment, we explore the **Router Mode** pattern, a configuration where a central team detects **context** (like language or domain) and routes queries to the right agent accordingly.
+
+> **What exactly is "context"?**\
+> In Agentic AI, *context* refers to the key details in a user's input that help the system understand how to respond appropriately. Think of it like clues that tell the AI what kind of help is needed. For example, if a user submits a query in Hindi, the language itself becomes part of the context. Similarly, if the query mentions "insurance claims" or "server configuration," that reveals the domain or topic the user is focused on.
+>
+> **Simple Analogy:**\
+> Imagine walking into a large helpdesk at an international airport. You speak to a receptionist and ask for assistance. Based on your language, destination, or issue (lost luggage vs. visa questions), they direct you to the right expert—someone who speaks your language or understands your problem. That receptionist is acting like the router agent. They’re not solving the issue themselves but are smart enough to know *who* should help you based on *context*. That’s exactly what the Router Mode does in Agentic AI.
+
+This method is especially effective in scenarios requiring multilingual or domain-specific support. Using the **AGNO framework**, we’ll see how to construct a language-routing team that handles diverse user inputs with precision and fallback logic—making it especially friendly for no-code or low-code setups.
+
+[Inspired by my Medium post.](https://dineshr1493.medium.com/all-you-need-to-know-about-the-evolution-of-generative-ai-to-agentic-ai-part-6-agentic-ai-a-39714050857b)
+
+## Router Mode: What it is?
+
+In Router Mode, rather than executing tasks itself, the team acts like a switchboard, rather than executing tasks itself. Its core responsibility is to:
+
+* Analyze user input
+* Detect the appropriate context (e.g., language)
+* Route the request to a specialized agent
+* Handle unsupported inputs gracefully
+
+<center><img src="/img/screenshot-2025-07-21-at-3.38.34 pm.png" width="600" height="550" alt="Route Mode" title="Route Mode"></center>
+
+### Use case: Multilingual chat support
+
+Imagine a chatbot that receives queries in different languages. Router Mode enables:
+
+* Language detection
+* Delegation to language-specific agents (e.g., Japanese, French, German)
+* Fallback messages for unsupported languages
+
+### Implementation: AGNO framework setup
+
+We’ll define a set of language-specific agents and create a routing team that delegates accordingly.
+
+#### Step 1: Define language agents
+
+```python
+from agno.agent import Agent
+from agno.models.anthropic import Claude
+from agno.models.deepseek import DeepSeek
+from agno.models.mistral.mistral import MistralChat
+from agno.models.openai import OpenAIChat
+
+english_agent = Agent(
+    name="English Agent",
+    role="You can only answer in English",
+    model=OpenAIChat(id="gpt-4.5-preview"),
+    instructions=["You must only respond in English"],
+)
+
+japanese_agent = Agent(
+    name="Japanese Agent",
+    role="You can only answer in Japanese",
+    model=DeepSeek(id="deepseek-chat"),
+    instructions=["You must only respond in Japanese"],
+)
+
+chinese_agent = Agent(
+    name="Chinese Agent",
+    role="You can only answer in Chinese",
+    model=DeepSeek(id="deepseek-chat"),
+    instructions=["You must only respond in Chinese"],
+)
+
+spanish_agent = Agent(
+    name="Spanish Agent",
+    role="You can only answer in Spanish",
+    model=OpenAIChat(id="gpt-4.5-preview"),
+    instructions=["You must only respond in Spanish"],
+)
+
+french_agent = Agent(
+    name="French Agent",
+    role="You can only answer in French",
+    model=MistralChat(id="mistral-large-latest"),
+    instructions=["You must only respond in French"],
+)
+
+german_agent = Agent(
+    name="German Agent",
+    role="You can only answer in German",
+    model=Claude("claude-3-5-sonnet-20241022"),
+    instructions=["You must only respond in German"],
+)
+ 
+
+ 
+```
+
+### Step 2: Create the Router team
+
+```python
+from agno.team.team import Team
+
+multi_language_team = Team(
+    name="Multi Language Team",
+    mode="route",
+    model=OpenAIChat("gpt-4.5-preview"),
+    members=[
+        english_agent,
+        spanish_agent,
+        japanese_agent,
+        french_agent,
+        german_agent,
+        chinese_agent,
+    ],
+    show_tool_calls=True,
+    markdown=True,
+    show_members_responses=True,
+    instructions=[
+        "You are a language router that directs questions to the appropriate language agent.",
+        "If the user asks in a language whose agent is not a team member, respond in English with:",
+        "'I can only answer in the following languages: English, Spanish, Japanese, French and German. Please ask your question in one of these languages.'",
+        "Always check the language of the user's input before routing to an agent.",
+        "For unsupported languages like Italian, respond in English with the above message.",
+    ]
+)
+```
+
+### Step 3: Run multilingual examples
+
+```python
+multi_language_team.print_response("How are you?", stream=True)         # English
+multi_language_team.print_response("你好吗？", stream=True)              # Chinese
+multi_language_team.print_response("お元気ですか?", stream=True)         # Japanese
+multi_language_team.print_response("Comment allez-vous?", stream=True) # French
+```
+
+### Output:
+
+```python
+[English Agent]: I'm doing great! How can I help you today?
+[Chinese Agent]: 我很好，谢谢你的关心。你呢？
+[Japanese Agent]: 元気です。あなたはお元気ですか？
+[French Agent]: Je vais bien, merci. Et vous ?
+```
+
+<center><img src="/img/screenshot-2025-07-21-at-3.38.47 pm.png" width="600" height="550" alt="Agent Mode Parameters" title="Agent Mode Parameters"></center>
+
+## Key parameters explained
+
+<table border="1" cellpadding="8" cellspacing="0" style="border-collapse: collapse; width: 100%;">
+  <thead style="background-color:#f2f2f2">
+    <tr>
+      <th>Parameter</th>
+      <th>Description</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td><code>mode="route"</code></td>
+      <td>Instructs the team to act as a switchboard, not an executor</td>
+    </tr>
+    <tr>
+      <td><code>show_members_responses=True</code></td>
+      <td>Displays individual agent replies for traceability</td>
+    </tr>
+    <tr>
+      <td><code>instructions\\\[]</code></td>
+      <td>Core router logic: detect language, enforce exclusivity, manage fallbacks</td>
+    </tr>
+    <tr>
+      <td><code>model=OpenAIChat(...)</code></td>
+      <td>Backbone LLM used by the router team for input analysis</td>
+    </tr>
+  </tbody>
+</table>
+
+## Why Router Mode works
+
+* Context-awareness: Inputs are analyzed for language, not just keywords
+* Agent exclusivity: Each agent strictly operates in its assigned language
+* Fallback resilience: Unsupported queries are met with a clear, unified message
+* Modularity: Each language agent is replaceable or extendable
+
+## Conclusion
+
+Router Mode in Agentic AI introduces scalable intelligence by structuring agents like a multilingual team. Rather than overwhelming a single agent, you delegate responsibility across specialists and keep interactions clean, accurate, and context-driven.
+
+With the AGNO framework, creating such intelligent, language-aware teams becomes seamless — and your agents become not just reactive, but well-organized and self-aware of their boundaries.
+
+A structured team, strict instructions, and intelligent routing — that’s the future of responsive AI.