After NextRise 2026: the questions that changed

Over 300 people visited the booth across two days. Teams already running AI agents came to ask how to secure them. The market has moved.

27 June 2026

At last year's NextRise, the questions were about access. This year, they were about control. Teams already running AI agents came to ask how to secure them. More than 300 people visited the booth over two days. Those conversations confirmed one thing: the market's understanding has caught up.

Most people who stopped at the booth fell into one of two patterns: they read the text on the banner and paused, or they saw the demo running on the screen and walked in. Once the explanation started, visitors would nod before we finished. They already understood what AI agents were doing in production environments—and went straight to "how do you control it?"

Every conversation over those two days pointed in the same direction. The market understands the problem now.

The AI+ zone

NextRise 2026 was a different scale altogether. With OpenAI, NVIDIA, Palantir, Google, Anthropic, and Perplexity all in attendance, COEX had official stages running simultaneously across the venue. Wherever you looked in the AI+ zone, there was a crowd.

AlpacaX presented across both days. Each session opened with a real AI agent security incident, then moved to Alpacon's execution control layer—about ten minutes, short and substantive. Each session moved directly into the live demo.

The mix of visitors was broad. Engineers and researchers from large enterprises had already deployed agents, or were close to it. Many were using their own control approaches but weren't confident in them. "We've already built out our agent infrastructure—how do we connect Alpacon to it?" was a question we heard repeatedly. Startup founders and developers were already running agents. Most recognized the risks but hadn't found a reliable approach; after hearing the explanation, many wanted to try it on the spot or said they planned to adopt it.

Investors came by steadily too. NextRise had set up dedicated 1:1 investor meetups, and interest in agent security infrastructure moved from booth visits into substantive conversations. Many visitors weren't infrastructure leads, but they brought up their own companies' problems and asked product questions anyway. The questions sounded different from person to person—but they shared one thread: "how do you control what an agent actually executes?"

Beyond the prepared mission

The booth had a live mission as well. Visitors logged directly into the Alpacon CLI and gave commands to an AI agent: check staging server health, detect log anomalies, attempt a blocked command. Visitors could try it themselves—seeing exactly what Alpacon intercepts and blocks.

The response went past what we prepared for. People who finished the mission didn't stop. They started typing commands directly into the CLI. "Does this one get blocked?" "How does the agent decide in this case?" It stopped being a mission and became exploration—visitors testing the boundaries of Alpacon's execution control on their own.

Visitors running the live CLI mission at the AlpacaX booth

We gave out gifts for completing the mission, but most people were more interested in Alpacon than the gifts. They wanted to know what gets blocked and what gets through—where the decision line sits. The moment that got the biggest reaction: commands streaming to the screen in real time. The moment an AI agent issued a command, it appeared live—and watching an out-of-scope command get blocked on screen was when "this actually works" landed.

KAIST Startup Scaleup Summit 2026

On the first day, KAIST Startup Scaleup Summit 2026 ran as a partner event in the same venue. It brought together startups from the Korea Advanced Institute of Science and Technology (KAIST) Startup Town Programme (KSTP), and AlpacaX was among them.

CEO Eunyoung Jeong took the stage to introduce AlpacaX—not just the company and product, but the context: the AI-native PAM category is forming now, and we're the team that has been building the execution control layer from the ground up. At a moment when AI agent governance is emerging fast as an investment category, the narrative connected technical foundation to market timing. After the session came 1:1 meetings with investors and enterprise leads—and some of them made their way to our booth afterward.

AlpacaX CEO Eunyoung Jeong presenting at KAIST Startup Scaleup Summit 2026

The market is changing its questions

Visitors at the booth had already identified the problem. "Once you give an agent permission, how do you control what happens after that?" was a specific question—and they wanted to understand what structure actually solves it.

In security product sales, this is decisive. It's hard to sell a solution to a team that hasn't recognized the problem. That prerequisite was already in place at this event. The quality of the questions was different too. Not "what is this?" but "how do we connect it to our environment?", "does it work when we have multiple agents?"—questions from teams already planning to adopt.

There was friction too: "We're using AI to make work easier—if there's this much human involvement required, does the point get lost?" That's a question only someone who has understood the product can ask. That tension is actually a good sign. Teams don't start calculating operational cost until they've figured out what the solution is.

The signal we read at this event is clear. Questions reaching this level mean the market is already moving. This category is forming now, and the team that builds the deepest execution control layer first defines it. The conversations at NextRise 2026 were two days of confirming that timing.