AgentX 2.0 - Build your own cross-vendor multi-agent AI team

@sebastian_mul  Combining the latest creations, I'll learn from you

@maali_baali You’re totally right! We’re working on making the Agent Builder much easier to use, including templates for common workflows.

In the meantime, a good way to get started is to copy any public agent from the homepage into your own workspace and tweak it to fit your needs. Just click the “Copy” button in the upper right corner of the agent page.

Want a quick guide on building your first AI agent? Check out this article and the video:

👉https://www.agentx.so/agentx-tutorial/how-to-build-an-ai-agent-beginners-guide

@kate_ramakaieva Hey! Thank you! Sure - you can customize your AI agents to match your brand, including your name, colors, logo, and website. This applies to both the chat interface and email notifications, so everything feels like a seamless part of your business. You can set this up easily in your AgentX dashboard under deployment settings.

For Enterprise plans, we also offer extra customization options if you need something more advanced.
Once Again, Thanks for your support!

@sidraarifali Thanks for the love! 🙌 AgentX is designed to scale seamlessly when running multiple agents across different channels simultaneously. Our platform can handle a high volume of concurrent conversations without lag, ensuring smooth and responsive interactions whether your agents are on WhatsApp, Slack, or your website.

Best practices to keep things smooth include:

• Organizing agents into specialized teams or MultiWorkforce setups to delegate tasks efficiently.

• Monitoring credit usage and setting up auto top-ups to avoid interruptions.

• Regularly updating knowledge bases to keep agents accurate and relevant.

• Picking the right LLM adapted to the tasks (using light and fast where possible)
  

Here you can find some information about the LLMs and picking the best one:
https://www.agentx.so/post/How-AgentX-Charges-for-AI-Usage

@igorsorokinua 
Thanks! Appreciate it 🙌

Auto-RAG in AgentX goes pretty deep - it doesn’t just pull in raw documents. It chunks content intelligently, understands structure (like Notion blocks or Google Doc headings), and can prioritize fresher or more relevant info depending on the context of the question. Of course, you are always able to guiderail your Agent in instructions.

So when your agent is answering something, it’s not just reading blindly, it’s choosing what matters.

More on how it works with Notion sync via Google Drive here:

👉 https://www.agentx.so/post/connect-agentx-ai-agent-with-notion

@evgenii_zaitsev1 Hey, thank you!
AgentX is built to slot right into the systems businesses already use: whether it’s Slack, WhatsApp, Google Drive, HubSpot, or custom platforms via API, Zapier, or MCP servers. Integration is fast and flexible, and we’re rolling out new connectors every week to keep pace with evolving tools.

We see companies using AgentX to deploy AI agents across support, sales, operations, and internal knowledge - each tailored with the right data, tools, and tone. It’s not just about automation, it’s about augmenting teams with AI that understands context and scales with your needs.

As AI continues reshaping how businesses operate, AgentX becomes the layer that connects people, platforms, and workflows, seamlessly.

Do you have a specific integration or use case in mind? I’d be happy to dive into it 💪

@rocsheh Thank you so much Zepeng for your kind word and feedbacks.

The history comparison between different deployment sounds like a useful feature! I will definitely let my team know. I believe we can absolutely enrich our Statistics page with feature. Will keep you posted!

@muhammad_adnan45 Awesome to hear! 🙌 Looking forward to seeing your own agent workforce in action - good luck, have fun, and reach out anytime you need. We’ve got your back! 🚀

@daniel_henry4 Thanks, we really appreciate that. You put it perfectly: it’s not just about giving people powerful tools, it’s about making them work together in a smart way. That shift opens up a whole new kind of usefulness.

@hehe_z Thank you Helena. We are working hard on provide the best service on MCP infra and also open sourced MCP server list.

Checkout: https://www.agentx.so/mcp and https://github.com/AgentX-ai/AgentX-mcp-servers

@rohanrecommends  @marcin_michalak  @gaia_chen24  @michael_talreja Awesome! Michael. That means a lot to us.

@lukas_timm Hi Lukas! It is definitely possible. In fact, we have already been supporting 10+ AWS MCP servers which allows our Agent to have certain control and execution ability with the system. And we are continuing to improve and adding more open sourced MCPs.

@monia_dohas Hi Monia, we currently do not have export, but we will support it once the Agent protocol is getting more mature and we know what format to export. We are actively work on the integration and keep the platform up to date with the newest tech requirement in AI Agent world.
Before that, you can always use our Python or JS SDK to connect the Agent with other project.

@monia_dohas Thanks so much! Really glad you’re enjoying the drag-and-drop builder.

Right now, you can move agents between workspaces and collaborate with teammates easily. We’re also rolling out a duplication feature next week, so you’ll be able to reuse agents across projects. On top of that, we’re working on letting you add knowledge directly to a workspace, so all agents there can share it.

And when it comes to multi-workforce teams - you can absolutely reuse agents and assign them to different teams. Each agent, along with its knowledge, can be shared across multiple workforces and work with different teams simultaneously.

@jackpark Thanks, Jack! We'd love to stay in touch - feel free to join our newsletter. In the meantime, have fun experimenting with your agents, we’re excited to hear your feedback! :)

@jackpark Remember to check out all the MCPs: https://www.agentx.so/mcp !

@tomik99 Thank you so much! 🙏

Really appreciate the kind words, they mean a lot to the whole team. We’ve worked hard to make AgentX feel like something you can use from day one, not just another AI wrapper. So it’s incredibly rewarding to hear that the clean execution and practical value are coming through.

Can’t wait to show you what’s next... ;) Onwards! 🚀

@patpijanowski Hey, thank you!

AgentX agents are highly customizable for any role or task! 🎯

• Custom Roles: Create multiple agents with unique functions (e.g., planner, support rep, sales bot).

• Tailored Instructions: Set detailed workflows and behaviors for each.

• Smart Integrations: Equip agents with the tools and data they need—CRMs, APIs, search, Google Maps MCP, and more.

• Agent Teams: Build multi-agent teams with a manager delegating tasks to specialists.

• Brand & UI Control: On Pro and Scale plans, fully customize look, feel, and branding.

• Easy Refinement: Edit anytime to improve or pivot agent performance.

@angelawang1605 Thank you for your question, Angela! And yes, that's totally possible!

AgentX is designed to plug into custom RAG setups, including MCPs and tools, so you can route part of your logic through your app/API/MCP and orchestrate other parts through the AgentX workflow layer.

You can also call out to external APIs or even trigger custom functions from within agents - happy to dive deeper if you’d like!

@angelawang1605 Great point Angela! We are definitely planning to accept building team of agents from outside AI agent routes. We are working closely on agent to agent protocol so that in future, your team of agents not only can be from different LLM vendor, but also from different platform source.

@joanna_polkowska Thank you! AgentX’s lead generation works by engaging visitors in natural, conversational dialogues rather than boring forms. Your AI agent asks questions like a human would to collect lead info such as name, email, phone, company, and more. Once collected, leads trigger notifications to you and confirmation emails to the visitor.
And yes, you can integrate lead generation directly with your CRM. AgentX supports seamless integration with popular CRMs including HubSpot, Wix CRM, and through Zapier, many others. For example, connecting to HubSpot CRM involves enabling lead generation in your agent’s Actions tab and adding your HubSpot API token as an endpoint - then leads flow straight into your CRM contact list in real time.

Here are some related posts:

https://www.agentx.so/post/how-to-create-a-lead-generation-ai-agent-chatbot-on-agentx-with-zapier-integration

https://www.agentx.so/agentx-tutorial/how-to-generate-lead-and-send-to-hubspot-crm

@ytk141 Thanks, Joe!

@ogucukavci You're absolutely right about the Featured Teams section - it will definitely continue to grow. We’re also considering highlighting community-created Workforces in the future, so users can share and discover real-world examples from each other. Great minds think alike 😊

We appreciate your suggestions on UI flow improvements - especially around shortcuts in the agent creation process. That’s really helpful input and we’ll definitely take it into account as we refine the experience.

Regarding the similarity in responses between Workforces and individual Agents - that’s largely driven by the underlying prompts, as you mentioned. With more advanced and varied prompt setups, we expect those differences to become more pronounced.

And of course, if you ever decide to explore creating your own Agents or Workforces and need a hand, don’t hesitate to reach out. I’d be more than happy to help you along the way!