TG-Staff Multi-Agent Chat Configuration Guide: Agent Count, Project Permissions, and Reception Limits

When your Telegram community or Bot customer service grows from a few user inquiries to dozens or even hundreds of messages per day, a single-agent model quickly becomes a bottleneck. One agent handling multiple conversations simultaneously leads to slower response times, longer user wait times, and near-impossible collaboration—a common challenge for every B2B SaaS team and Telegram operator.

TG-Staff’s multi-agent chat feature is designed to solve this problem. It allows multiple agents to log into the web console simultaneously, independently handle different Telegram user conversations, and ensure timely responses through routing rules, project permissions, and reception limits. This article will guide you step by step through configuring multi-agent chat, from choosing the number of agents to setting project permissions, and managing routing rules and reception limits, helping your team quickly build an efficient online customer service system.

Why Configure Multi-Agent Chat—Bottlenecks of the Single-Agent Model

Before diving into configuration, let’s clarify a key question: Why isn’t the single-agent model sufficient?

Imagine you run a cross-border e-commerce Telegram community with 50 users daily inquiring about order status, logistics, and return policies via a Bot. If you have only one agent, they would need to:

• Reply to each user sequentially, unable to handle multiple conversations simultaneously
• While replying to A, users B and C wait, with response times potentially exceeding 10 minutes
• No division of labor, such as having one agent handle logistics and another handle after-sales
• Once the agent goes offline or takes a break, all inquiries pile up

The result is decreased user satisfaction and potential loss of orders. Multi-agent chat’s core value lies in expanding customer service capacity from a single person to a team, achieving scaled reception through parallel processing, permission isolation, and intelligent routing.

TG-Staff provides a complete multi-agent chat solution that doesn’t require development from scratch—just configure it directly in the web console. Let’s break down its core concepts.

Core Concepts of TG-Staff Multi-Agent Chat: Agents, Sessions, and Projects

To configure multi-agent chat correctly, you need to understand three foundational concepts—agents, sessions, and projects—and how they work together.

Agent (Staff Seat)—Independent Customer Service Workstation

Each agent represents an independent login account for a customer service representative. After logging into the web console (https://app.tg-staff.com/), agents can see their assigned session list and reply to Telegram users in real time.

• Agent quota: Varies by plan. The Standard plan (approx. 8.99/month) offers 3 agent seats; the Pro plan (approx.16.99/month) offers 20 agent seats. A 3-day free trial lets you experience all features.
• Independent operation: Each agent has their own login credentials, non-interfering. Agents can transfer sessions, add private notes (Pro plan), enabling collaboration without conflict.
• Permission control: Restrict which sessions an agent can see at the project level (detailed below).

Session—Channel from User to Agent

When a Telegram user first interacts with your Bot (e.g., sending the /start command or clicking a menu button), TG-Staff automatically creates a session for that user. This session is the communication channel between the user and the customer service team.

• One-to-one mapping: Each Telegram user corresponds to one session. User messages enter the session, and agent replies are sent back to the user through the session.
• Parallel multi-session: In multi-agent mode, multiple user sessions can be handled simultaneously by different agents. For example, Agent A handles User X’s order issue while Agent B handles User Y’s logistics inquiry, without interference.
• Session status: Sessions can be in “Pending,” “In Progress,” or “Closed” states, helping agents prioritize.

Project—Binding Bot and Customer Service Team

A project corresponds to a Telegram Bot (connected to TG-Staff via Bot Token) and its associated customer service team configuration. You can create multiple projects in TG-Staff, each binding a different Bot.

• Multi-project support: Plans support different numbers of Bot projects. For example, you could create separate projects for “Main Community Bot,” “After-Sales Bot,” and “Event Bot,” each managed independently.
• Project-to-agent mapping: Configure the “Project Customer Service Scope” to decide which agents can handle sessions under that project.

You can think of these three concepts as: Project is the container, binding the Bot and defining service rules; Agent is the operator, logging into the web console to serve users; Session is the real-time communication channel between user and agent. Multi-agent chat enables multiple agents to handle multiple sessions simultaneously within one or multiple projects.

Step 1: Choose a Plan and Configure Agent Count

The first step in configuring multi-agent chat is ensuring your plan supports enough agent seats. TG-Staff’s plans cater to different team sizes:

Plan	Agent Seats	Suitable For	Monthly Fee Reference
Free Trial	All features	3-day trial period	Free
Standard	3 agents	Small teams, 1-3 people on shift	Approx. $8.99/month
Pro	20 agents	Medium to large customer service teams, multi-project operations	Approx. $16.99/month

Agent seats are the maximum number of agents online simultaneously, not the total team size. For example, if your team has 5 people but scheduling ensures at most 3 are online at once, the Standard plan’s 3 agent seats suffice. If you need 4-5 agents online simultaneously, consider the Pro plan.

Operation Steps:

1. Log in to https://app.tg-staff.com/ and enter the console.
2. On the “Settings” or “Plan” page, view your current plan and upgrade as needed.
3. Create agent accounts: In “Agent Management,” add the agent’s name, email, and password. Each agent will receive a unique login link.

Step 2: Project Permission Configuration — Who Can Handle Which Conversations

Once the number of agents is determined, the next step is to configure the scope of customer service for each project. This step determines which agents can handle conversations for specific Bot projects. TG-Staff offers two modes: All Agents and Specified Agents.

All Agents Mode — Suitable for Single-Project Teams

If your team operates only one Telegram Bot project, or you want all agents to see all conversations, choose the “All Agents” mode.

• Features: All agents (within permission scope) can see all conversations under the project and participate in handling them.
• Use Cases: Single community management, customer service team for a single product line.
• Configuration: In the project settings, set “Project Agent Scope” to “All Agents.” After logging in, agents will see all pending conversations for the project.

Specified Agents Mode — Suitable for Multi-Project Isolation

If your team operates multiple Telegram Bot projects simultaneously (e.g., multiple product lines, multiple communities, or bots in different languages), you may want different agents to handle only specific projects to avoid data confusion or permission breaches.

• Features: Only authorized agents can see and handle conversations for specified projects. Other agents cannot see the project at all.
• Use Cases:
  • Different product lines handled by separate customer service teams
  • Different language communities maintained by agents proficient in those languages
  • Isolation of customer service permissions between internal and external bots
• Configuration: In the project settings, set “Project Agent Scope” to “Specified Agents,” then select the authorized agents from the agent list.

Example: Suppose you have two projects — “Main Store Bot” and “After-Sales Bot.” You can assign Agents A and B to “Main Store Bot” and Agents C and D to “After-Sales Bot.” After logging in, Agent A only sees conversations from the main store, and Agent C only sees after-sales conversations, achieving complete isolation.

Step 3: Set Conversation Distribution Rules and Handling Limits

After configuring agents and projects, you need to decide how new conversations are assigned to agents. TG-Staff offers two distribution rules: Round Robin (default) and Online First.

Round Robin

The default distribution method. The system assigns new conversations to authorized agents in a fixed order, rotating sequentially.

• How It Works: Suppose agents A, B, and C (all authorized). New conversation 1 is assigned to A, conversation 2 to B, conversation 3 to C, conversation 4 to A again, and so on.
• Suitable Scenarios: All agents have fixed schedules and are online during the same hours. For example, a customer service team is fully online from 9:00 AM to 6:00 PM. Round robin ensures fairness.
• Note: If an agent is offline, the system skips them and continues to the next online agent.

Online First

New conversations are prioritized for agents currently online. When all agents are offline, it falls back to round robin.

• How It Works: When a new conversation arrives, the system checks the online status of all authorized agents and assigns it randomly or sequentially to an online agent. If all agents are offline, conversations are queued following round robin rules, but agents cannot respond immediately.
• Suitable Scenarios: Agents have irregular online times, such as remote teams or part-time customer service. Online First maximizes the use of online resources and reduces user wait time.
• Note: If an agent is online but handling many conversations, the system will consider the handling limit (see below) to avoid overload.

Reception Limit – Preventing Overload of Individual Agents

In addition to routing rules, you can set a reception limit for each agent, i.e., the maximum number of concurrent conversations they can handle. When an agent reaches their limit, new conversations will not be assigned to them but to other available agents.

• Why it’s needed: Prevents an agent from being overloaded with too many conversations, which could lead to delayed responses or reduced service quality.
• How to set: In the agent management page, set a “Max Concurrent Conversations” for each agent (e.g., 5 or 10). Adjust based on the agent’s efficiency and business complexity.
• Best practice: New agents can start with a lower limit (e.g., 3) and gradually increase as they become more proficient. For complex inquiries (e.g., technical issues), set a lower limit to ensure response quality.

Configuration Steps Summary:

1. In project settings, select a routing rule (Round Robin or Online First).
2. In the agent management page, set a reception limit for each agent.
3. Save the configuration, and new conversations will be automatically assigned according to the rules.

Diversion Link – Multi-Channel Traffic Routing and Attribution

Multi-agent conversations not only solve internal reception efficiency but also integrate with external traffic routing. The Diversion Link is a feature available in TG-Staff Standard and above plans, designed to precisely route traffic from ads, social media, or other channels to your Telegram Bot. Combined with conversation routing, it creates a complete funnel from traffic generation to human agent handling.

A Diversion Link is a short URL under the TG-Staff official domain (e.g., https://app.tg-staff.com/{code}). When a user clicks this link, the system captures the visitor’s IP address, browser information, and custom parameters carried in the URL (such as utm_source, campaign_id) before redirecting to your Telegram Bot. This information can be used for:

• Ad attribution: Track which channel (e.g., Facebook ads, Google search, KOL promotion) brought in the inquiry volume.
• Multi-channel tracking: Generate different Diversion Links for different campaigns to analyze conversion performance.
• User profile enrichment: Captured information can be associated with the user session, helping agents understand the user’s source before the conversation.

Combined with the routing rules configured earlier, user sessions from different Diversion Links are automatically assigned to authorized agents, achieving seamless integration from ad click to human handling. For teams in Web3, cryptocurrency, or outbound marketing that require precise traffic attribution, Diversion Links are a crucial tool for improving ROI.

Best Practices: Daily Operational Tips for Multi-Agent Conversations

After configuration, daily operations are equally important. Here are some actionable tips:

1. Agent Handoff Process: When an agent needs to transfer a conversation to a colleague, use TG-Staff’s “Transfer Conversation” feature. You can add a note to inform the receiving agent of the current progress. The Professional plan also supports private notes, allowing agents to record key information internally without affecting the user’s view.

2. Peak Hour Routing Strategy: If your business has clear peak inquiry periods (e.g., after promotions, 10:00-12:00 on weekdays), configure the routing rule to “Online First” in advance and ensure enough agents are online during peak times. If agents are insufficient, consider using TG-Staff’s Bot auto-reply feature (visual command flow) to inform users of waiting times or guide them to self-service.

3. Agent Training Points: Ensure every agent is familiar with the Web console operations, including:

   • How to view the conversation list and mark conversation status (Pending/In Progress/Closed)
   • How to use the auto-translation feature (Standard plan includes AI translation; Professional plan supports Google and DeepL professional translation)
   • How to view user profiles (Professional plan) to quickly understand user history and behavior
   • How to follow content moderation rules (Professional plan) to avoid sending sensitive information

4. Leverage User Profiles to Improve Service Quality: The Professional plan’s user profile feature allows agents to view historical conversation records, tags, source channels, etc. Agents can review the profile before handling a conversation to quickly understand the user’s background, reduce repetitive questions, and improve first response quality.

Frequently Asked Questions

Q: How many agents can be online simultaneously in the Standard plan?
A: The Standard plan (approx. 8.99/month) supports 3 agent seats, meaning up to 3 people can log into the Web console simultaneously to handle Telegram users. The Professional plan (approx.16.99/month) supports 20 agent seats, suitable for medium to large customer service teams. See the official pricing page for details.

Q: Can I assign different customer service teams to different Telegram Bot projects?
A: Yes. In TG-Staff’s project permission settings, you can choose “All Agents” or “Specified Agents” for each project. If you select specified agents, only authorized agents can view and handle conversations under that project, achieving multi-project data isolation. For example, for a “After-Sales Bot” project, only authorize the after-sales team agents; other agents cannot view it.

Q: What is the difference between Round Robin and Online First routing rules?
A: Round Robin assigns new conversations to authorized agents in a fixed order, suitable for fixed-schedule teams; Online First prioritizes agents who are currently online, and falls back to Round Robin when all agents are offline. It is recommended for teams with irregular online hours to use Online First mode to maximize online customer service resources.

Q: Which plan includes the Diversion Link feature?
A: The Diversion Link is a feature available in the Standard plan and above. It generates a short URL under the TG-Staff official domain (e.g., https://app.tg-staff.com/{code}) that captures visitor information (IP, browser, URL parameters) before redirecting to your Telegram Bot, used for ad attribution and multi-channel tracking. The Standard plan includes this feature.

Q: What happens to new conversations if all agents are offline?
A: When all agents are offline, new conversations enter a waiting queue. The “Online First” routing mode automatically falls back to Round Robin, but note that agents are not online to respond immediately. It is recommended to set up Bot auto-reply to inform users of wait times or leave contact information. When agents come online, conversations in the waiting queue will be reassigned according to the rules.

---

If you are looking for a tool to quickly build a multi-agent conversation system, try TG-Staff. It provides comprehensive capabilities from agent management, project permissions, routing rules to traffic attribution, helping your Telegram customer service team operate efficiently.

• Start Free Trial: Visit https://app.tg-staff.com/ to experience all features for 3 days.
• View Documentation: Go to https://docs.tg-staff.com/ for operation guides and API reference.
• Contact Support: For any questions, directly contact the official Bot @tgstaff_robot for one-on-one support.