Ever feel like you're one subnet away from passing the CCNA? You've watched the videos, read the guides, and aced the flashcards-but when that timer starts, everything blurs into a VLAN of confusion. If you've ever shouted, "There's got to be a better way!"-you're not alone.
Welcome to the "Almost There" Club-a population of way too many. Here's the good news: Learn with Cisco is here to help. And this iteration of The AI Break will explain how to get AI to help you push across that CCNA finish line.
Here's the secret formula:
- Spin up CCNA lablets in Cisco Modeling Labs with the help of AI
- Use our CML MCP Server to learn how to let AI help you create your lablets.
Bring in AI as your personal lab coach.
Think of it as a Peloton for your network skills-except instead of spinning wheels, we're spinning up topologies.
So, why does this matter, anyway?
Let's be honest-most CCNA candidates don't fail because they didn't study. They fail because they didn't practice the right way. Reading about OSPF or VLANs is one thing; troubleshooting them under pressure is another. The CCNA certification exam tests how well you can think and act like a network engineer, not just recall facts.
AI, Model Context Protocol (MCP), and Cisco Modeling Labs (CML) to the rescue
As Joe Clarke highlighted in his recent blog, Speak Your Lab into Existence with AI-Driven Cisco Modeling Labs and MCP, by combining Model Context Protocol with CML, engineers can create AI-aware topologies that understand context and can even self-describe their structure.
That's where this approach changes everything. When you combine AI, Model Context Protocol (MCP), and Cisco Modeling Labs (CML), you're creating an environment that's dynamic, hands-on, and guided:
- Cisco Modeling Labs (CML) gives you the sandbox to build and break networks safely.
- Model Context Protocol (MCP) connects that environment to an agentic AI layer that understands your topology.
- AI becomes your mentor that can generate tasks mapped to the CCNA blueprint, create mini challenges, ask you to verify routing tables, and even nudge you with hints when you're stuck.
This kind of hands-on repetition builds confidence, muscle memory, and the intuition that separates a pass from a near miss.
I need to study. Can you just get to it?
Now that we know why this matters, let's get practical.
Here's what you'll need to get started:
- CML MCP Server (setup instructions)
- CML installed (local installation) -Free
- AI model configured - (Use your preferred model or agent that can talk to MCP). Here are some free options:
- Claude Desktop -Free
- LM Studio -Free local Models
Let's get AI to build the lablet
Now comes the fun part-letting AI do what it does best: translate intent into action. You don't need to memorize command syntax or stare at a blank topology wondering where to start. Just tell AI what you want to practice, and it builds the lab environment around that goal.
Say you're brushing up on VLANs and inter-VLAN routing. All you need to do is tell AI, "Build me a small lab to practice VLAN segmentation and routing between two subnets." Within seconds, the MCP-connected AI spins up a topology in Cisco Modeling Labs (CML): a couple of routers, switches, and hosts, pre-wired for your scenario.
We can even go one step further: Ask the AI to interpret the CCNA exam blueprint and generate labs that cover specific topics.
Here's a good starting prompt you can use:
Prompt >
I am a CCNA exam candidate, and I am building a hands-on discovery lab to practice for the exam using the CCNA blueprint. Here are the steps I want you to take: 1. Create a hands-on lab task based on the CCNA blueprint here: https://learningcontent.cisco.com/documents/marketing/exam-topics/200-301-CCNA-v1.1.pdf 2. Only have at most **One** related Task, provide enough information to complete the task without giving away the answers 3. Create a new CML lab for this lab and provide the instructions for what I need to perform as lab notes. Note: lab notes can make use of markdown for richer text. 4. Create the lab topology in CML using `CML Free version`: - Only Use `IOL`, `IOL-L2`, `ASAv` `Desktop`, `Server`, `unmanaged switches`, and `external connectors`. - Do NOT use `IOSv` or `IOSv-L2`. Note: unmanaged switches and external connectors do not count as nodes when considering the maximum of five. - Do not create the full topology right away. 5. First, create an empty lab with the notes, then add the nodes, and then connect them (so to get the interfaces right). 6. Be sure to update the lab notes with the new, correct interface names. 7. Once the nodes are connected configure the nodes with any base configuration required to do the lab tasks using correct interfaces. 8. Upon completion make sure you start the lab. - Start with Steps 1-4 then wait for my instructions to proceed - Next, do Steps 5-8
And just like that-you have a custom CCNA practice lab ready to go:
- A topology built to the exact specifications you need from the CCNA blueprint
- Lab notes that guide you through the tasks
Now, let's get the AI to validate and grade your work
AI built your lab, and you've completed the tasks. Now it's time to see how close your work is to the real thing.
With PyATS configured on the MCP server, the AI can read device configurations, verify results, and highlight where things don't line up.
Prompt >
ok! I've completed the tasks - I need you to check my config and validate I've completed the tasks correctly. - Grade my work, if I made a mistake let me know where I went wrong and why. - Do not make any changes to my config. - Provide me with next steps to correct my mistakes. - Suggest additional practice tasks to help me improve my skills based on my mistakes.
Clearly, I need more practice
Well, let's just say my first AI-graded lab didn't exactly earn me a gold star:
- I forgot to create VLAN10and VLAN20 on SW1
- I put PC1in VLAN20 instead of VLAN10
(Don't judge.) In short, I clearly need more practice.
Domo arigato, Mr. Roboto
And that's exactly the point. This whole setup is about creating a space where you can experiment, make mistakes, and learn faster. The AI gives you real-time feedback, showing not just what's wrong but why, and guiding you toward the correct fix.
Every iteration makes you better. Every mistake turns into progress. And before you know it, you'll start troubleshooting and configuring like it's second nature.
Keep practicing. Keep testing. Keep learning. Because the more you use this hands-on, AI-assisted approach, the more confident you'll be when you sit for your CCNA exam - and beyond.
Turn 'almost there' into 'certified'
And remember, this doesn't just stop at CCNA. The same method works beautifully for CCNP, DevNet (CCNA Automation), Cybersecurity, or even CCIE. Any certification that measures how you think through a problem benefits from having AI as your lab partner.
So go ahead. Fire up your MCP server, let AI generate your next challenge, and keep at it. Practice until your "almost there" becomes "certified."
Ready to schedule your 200-301 CCNA exam? Lock in a second chance at success with Cisco Exam Safeguard.
Read next:
Speak Your Lab into Existence with AI-Driven Cisco Modeling Labs and MCP
Sign up for Cisco U. | Join the Cisco Learning Network today for free.
Learn with Cisco
X | Threads | Facebook | LinkedIn | Instagram | YouTube
Use #CiscoU and #CiscoCert to join the conversation.