With 2025 wrapping up, I’ve had some time to reflect on life and the fortune I’ve had meeting so many great people in the tech community. I’ve been creating content (videos, blogs, etc.) for a few years now and have fell off quite a bit over the past couple years because of shifting my focus to more important life priorities, but I truly appreciate everyone in the community that stays tuned into my content, so thank you! Now let’s get into the blog!

Summary

This was my third AutoCon (attended AC 0 and 2) and, once again, I had another new experience while attending AutoCon. AutoCon 0 was the inaugural conference – the first conference focused on network automation. At AutoCon 2, I was selected as a speaker and it was my first time speaking in-person to a room of 500+ people. For AutoCon 4, I convinced my company to sponsor the conference and so I attended my first conference as a sponsor.

As a sponsor, I spent most of my time at our booth (as you can expect). I’ve always heard “booth duty” isn’t the most glamorous, and I had glimpses of what people meant, but for the most part, I enjoyed meeting everyone that approached our booth. Besides giving the general pitch about my company and the network automation consulting services we offer, I was also able to talk about the Cisco pyATS book I co-authored with my good friend John Capobianco because we were giving away a signed copy each day of the conference as part of a giveaway. There was also a good amount of people that approached the booth that either knew me from the content I published on my blog or on my YouTube channel, or through online interactions we’ve had over the years, which was really cool! Enough about me though, let’s dive into what you’re really here for – my experience at AutoCon 4.

Conference Highlights

As you can expect, AI was front and center in almost every vendor announcement, and was mentioned in a majority of the talks. There are plenty of other blogs from vendors and individuals that go into the details of every announcement made at the conference, but I wanted to touch on a couple announcements I found interesting.

Itential announced FlowAI, which essentially brings agentic AI capabilities to the Itential automation platform. I was able to see a demo of FlowAI from a couple folks at Itential and I can say I was thoroughly impressed. It wasn’t the magic of AI executing a task on it’s own that impressed me, but how well the agent’s “reasoning” was documented within the Itential platform. Traceability is huge when it comes to tracking context and how the agent “thought” about executing a task/workflow. FlowAI is still in private preview, but I can see the promise.

NetBox Labs had a few products and features on display, including NetBox Copilot, NetBox MCP server, NetBox Discovery, NetBox Assurance and their freshly announced product NetBox Observability. NetBox Observability intrigues me because it makes sense, but it also branches outside of only documenting the network. It makes sense because NetBox already has all your infrastructure data, including the intended state of your network, so having that additional context combined with real-time telemetry really provides a complete picture of how your network is running (operating state) vs how it should be (intended state).

Before jumping into my take about the conference, I did want to give kudos to the conference organizers and NAF decision makers for including multiple tracks at AutoCon 4, including an “Advanced” technical track and a “Leadership” track, along with the traditional “General” sessions. Beyond the tech, it’s important to shed light on the cultural and leadership buy-in required in the process of adopting network automation in an organization.

My Take

If AutoCon 4 was to answer NAF’s ultimate question, “Why haven’t we seen full adoption of network automation, yet?”, the answer I gathered from the conference would be “Because we didn’t have AI”. Obviously, that’s a tongue-in-cheek statement, but I know it generated some sort of emotional response from you :). With that being said, let’s dive into why that answer has some truth to it (based on my own experiences).

Ideation

“I don’t have time to focus on use cases. I’m always fighting fires”. Sadly, this is the reality for many network engineers. They are busy with their day job of keeping the network up by responding to incidents, answering requests, and delivering on projects. They don’t have the time to even think about what they want to automate – let alone try to build out a formal business case to present to management and other stakeholders to invest in the people, necessary tooling and training. This is a perfect use case for AI. I’m not saying AI should write your business cases (maybe?), but AI can be a helpful companion during the ideation phase of building out those business cases – identifying specific use cases, pros/cons of certain approaches, and even example workflows based on specific tooling. It’s not going to give you all the answers, but within minutes, AI will perform a lot of the Google legwork that you were planning to do during your “free time”. I’ll wrap up by providing the general AI disclaimer: Hallucinations are real. Don’t take all information generated by AI at face value. AI can expedite your business/use case research, but always double-check your work before submitting anything formal.

Proof-of-Value / MVP

“Vibe coding”. If you’ve never heard of it, I’ll save you a Google search or question to your favorite LLM. Vibe coding is the idea that AI generates and refines code at your direction using natural language. Essentially you tell AI what to build for you. There are many stories of frustration with vibe coding, boiling down to users telling the AI to “just fix it”, but vibe coding can actually be very useful for discovering “art of the possible”. AI can help create a full-stack application that you didn’t think was possible because maybe you don’t know frontend web development or databases. I’ve personally used it to build web apps with web frameworks I wasn’t familiar with so that I could learn some of the basics by referring to a project relative to my interests vs writing another “hello world” app where I copy code from a tutorial. Vibe coding may not produce what I’d consider a “production-ready” application out of the gate, but it can help bring your ideas to life, create a vision, and even help encourage buy-in from others.

Code Development

If you’re further in your automation journey and more of a “builder”, I’ve seen a lot of success with AI analyzing and providing recommendations when developing code. Not to get confused with vibe coding, I’m talking about using AI as more of a coding copilot that watches over your shoulder. It doesn’t rule out AI generating code on it’s own, but I want to be clear that a human is in the driver’s seat, with AI supplementing. This is more of an efficiency, but ultimately can evolve into agentic AI, where you can give tasks or issues to AI to figure out and ultimately develop a solution. This is where a solution like FlowAI from Itential fits into the picture – interprets user’s intent, compares different options, and ultimately performs deterministic execution. This is sometimes called “AI reasoning”. I think agentic AI is going to be a theme we see continue in 2026, as it ultimately makes our lives easier and adds bandwidth to our already small infrastructure teams.

Conclusion

I know I talked a lot about AI, but at the end of the day, AI is just another tool. As with any new tool, you need to validate the outcomes regardless of what’s promised in the hype. AI is here to stay, but it’s up to you on how it’s used in your life (personal and professional). Professionally, I think organizations will (if not already) overextend AI and what it can accomplish, which is common with any new tech revolution (remember how cloud would eliminate the need for private datacenters?). So, in short, I think AI will ultimately help increase network automation adoption, but it’s not a “silver bullet”. As I saw in a recent Claude commercial by Anthropic, “Keep Thinking”.

Thank you again for reading this post! As always, feel free to reach out to me on Twitter/X (@devnetdan) or on LinkedIn. Until next time!