From Listening to Doing: An Interactive ServiceNow AI Meetup in Hungary

[AttilaVarga]

Making the Meetup More Engaging

As one of the organizers of the Hungarian ServiceNow Developer meetup, I'm always thinking about what kind of content we should prepare from event to event. We really don't want, that a meetup to become boring, because that would quickly lead to fewer participants. For some time now, I have felt that we should bring in an interactive game, where attendees can actively take part.

I know from my own experience that if I only read about a topic, my knowledge stays shallow. But if I can try it myself, the learning becomes much stronger. I can better understand the topic, use it later in my work, and even build more knowledge on top of it when needed.

So the main goal was to encourage people to move from being passive listeners to becoming active participants.

AI Agents as the Main Topic

The main focus of our end-of-year meetup was the introduction of AI agents and Agentic Workflows.

We started from the basics and explained the building blocks: the ServiceNow core features that AI agents can use. These are the so-called tools. Then we talked about what an AI agent is, how it works, and finally how ServiceNow can coordinate (orchestrate) multiple AI agents to automate more complex solutions.

Related to this topic, I looked for a challenge that was interactive, fun, and also matched the Christmas atmosphere of the end of the year. For this, I reused and reshaped one of my earlier ideas. At that time, I used a Lego vehicle to show ITOM automation from a different point of view. (Is IT Operation Management able to control a Lego vehicle?)

Now, the goal was to make the AI solutions available in ServiceNow easier to understand and closer to people.

A Christmas-Themed Interactive Challenge

It was Christmas time, so I built a “motorized” sleigh for Santa. 🙂

 

The story was that the sleigh got stuck, and Christmas was in danger. We had to ask help from ServiceNow AI agents to remotely control the sleigh through a “dangerous” path in order to free the stuck sleigh and save Christmas.

How the Game Worked

I created a simple track. The task itself was very straightforward. We did not want to give a heavy technical or coding challenge, so I prepared all technical parts in advance:

• I created a REST API to control the Lego vehicle (Lego uses Bluetooth for controlling the vehicle, so the REST API server was running locally).

• To keep things simple, I added a distance sensor to the sleigh, which automatically stopped it before hitting anything.

• I configured a MID Server to handle the control, because the REST API is not accessible from the cloud.

• I built the required ServiceNow Flows to control the Lego vehicle (these were the AI Agent tools).

• I created AI agents responsible for controlling the sleigh:

  • Move forward (or backward) for a predefined time or continuously

  • Stop the vehicle

  • Turn the vehicle

  • Check whether the vehicle is moving or stopped

My first test execution during the preparation looked quite good:

• I also prepared a short guide explaining how the API works.

The task was to create an Agentic workflow, based on the presentation, which was the first part of the meetup event. Using the AI agents, the workflow had to control the sleigh through the path. In short, the challenge was about writing a good prompt. The gamification part was simple: the team whose sleigh finished the track in the shortest time won the prize. We rewarded the top two teams.

Because only a limited number of ServiceNow instances were available, we introduced a special feature called Developer Mode. When this mode was enabled, the REST API calls were mocked but returned valid responses. In this case, the vehicle did not actually move. This allowed all teams to develop and test their workflows at the same time on the same instance.

We planned a kind of dry run, but unfortunately there was not enough time for it. So every solution was tested "live". Looking back, this made the results even more fun. 🙂

Running the Competition

Four teams joined the challenge, with team sizes ranging from one to five people. It was great to see their enthusiasm. Many times they came to me with questions, but I always made sure to help only enough so that no team gained an unfair advantage.

Teams had about 40–50 minutes to write their prompts. After that, the competition started. We ran each workflow one by one and measured the time.

It was very exciting, we laughed a lot, and people really enjoyed it.

 

Results and Feedback

Overall, the meetup was very successful: a strong presentation in topic AI agents, followed by the game, and finally many great conversations. I am sure that participants and even those who did not play, left with a much clearer understanding of how AI agents and agentic workflows work, and with useful knowledge they could take home.

Lessons for Future Meetups

I will continue to look for similar topics for our future meetups. I strongly believe that this kind of approach not only helps knowledge sharing and learning, but also adds a playful challenge, which (based on our experience) motivates people better.

Closure

Of course I created a prototype during the implementation. If you would like to see how it works, check this video:

#aiagent #agentstudio #agenticworkflow #developermeetup