Skip to main content

Event Promotion

How to fill a room for your Applied AI event.


The Principle

Promotion is not a one-time announcement. It's consistent, repeated visibility in the weeks leading up to the event. The people who need to hear about your event are busy. They need to see it multiple times across multiple channels before they register.

The specific channels depend on your local context. A campus chapter promotes differently than a city chapter. But the underlying discipline is the same: start early, stay consistent, and use every channel available to you.


Timeline

WhenWhat
3-4 weeks beforeEvent listed on Luma/Meetup, initial announcement on socials
2-3 weeks beforeBegin consistent posting cadence (multiple times per week)
2 weeks beforeText your friends and personal contacts directly
1-2 weeks beforeAnnounce in lecture halls, Slack groups, email lists
Week ofFinal push: daily posts, direct messages, day-of reminders

Channels

Social Media (Your Personal Accounts)

This is the highest-leverage channel early on. The chapter leader's personal social presence drives most initial signups.

What works:

  • Post multiple times a week in the lead-up, not just once
  • Newsjack: tie the event to whatever's happening in AI that week. A new model drops, a company announces layoffs, a tool goes viral. Connect it back to why people should come to your event. AAS chapter leaders in Austin do this consistently.
  • Share behind-the-scenes prep (confirming speakers, setting up the venue, your own excitement)
  • Tag speakers and partners in posts for amplified reach
  • Post the day of the event with a “last chance to register” message

What doesn't work:

  • A single announcement post and nothing else
  • Generic “come to our event” language with no hook
  • Posting only on the event platform and expecting people to find it

Bold All-Caps Headlines

Every social post opens with a bold all-caps headline. This is a curiosity-driving hook, not a topic label. The goal is to make someone stop scrolling because they want to know more.

On X, use markdown bold: **HEADLINE HERE**

On LinkedIn, use Unicode bold characters (tools like YayText can generate these).

A good headline makes a claim, asks a provocative question, or teases a story. A bad headline just names the event.

  • Bad: “APPLIED AI LIVE #3 RECAP” (this is a label, not a hook)
  • Good: “WHY THE MAYOR OF AUSTIN AND I AGREE THAT AUSTIN SHOULD BE THE APPLIED AI CAPITAL OF THE WORLD”

The headline should work on its own, even if someone never reads the rest of the post. It should make them curious enough to keep reading.

Event Plug Placement

Where you put the event plug matters. Getting it wrong makes your post feel like an ad instead of a genuine share.

On X: Put the event plug in a reply to the main post, not in the post body. The main post should feel authentic and reflective. It should be about a real thought, an observation, something you learned, or something happening in AI. The reply handles the event plug, with @handles for speakers and the event link. This separation is important: people engage with the main post because it's interesting, and the reply catches the ones who want to take action.

On LinkedIn: The event plug goes inline at the end of the post. Keep it to a soft paragraph with speaker names, what they're presenting, venue, and the registration link. LinkedIn's format is more forgiving of longer posts with a call to action at the bottom.

Speaker Descriptions

When you mention speakers, don't just list names and titles. Write warm, specific descriptions of what each speaker is presenting and why it matters.

Compare these two approaches:

Generic: “Michael Daigler, Developer Advocate at Apify”

Specific: “Michael Daigler from Apify is breaking down OpenClaw x Apify for building business automation agents”

The specific version tells the reader what they'll actually learn. It gives them a reason to show up beyond name recognition. This is especially important when your speakers aren't yet well known. The description of what they're presenting does the work that their name can't do yet.

Every Post Is a Promotion Opportunity

Every social post you publish (even ones that aren't about the event) should include an event plug. On X, this goes in the reply. On LinkedIn, it goes at the end of the post.

The plug should feel natural, not forced. Something like: “If you want to be in rooms where these conversations happen in person, here's the next one.”

This works because you're already talking about AI topics that your audience cares about. The event plug is a logical next step, not a detour. The key is making the main post genuinely interesting on its own. The event plug is a bonus, not the point.


Direct Messages and Texts

The most effective promotion channel, especially for your first few events.

Text everyone you think would get value from the event. Personally. Not a mass blast. A real message:

Hey, I'm running an Applied AI event on [date] at [venue].
Practitioners sharing real case studies, live architecture
session, good networking. Think you'd dig it. Want to come?
[Luma link]

This feels like a lot of work. It is. It's also how you get 50-75 people in a room the first time. After your events have a track record, word of mouth carries more of the load.


Email Lists

If you have access to relevant email lists (university department lists, local tech community newsletters, coworking space announcements), use them.

For campus chapters:

  • Department email lists (CS, business, engineering, design)
  • Student organization newsletters
  • Faculty who are willing to forward to their classes

For city chapters:

  • Local tech community newsletters
  • Coworking space announcement boards
  • Partner organization email lists (with permission)

Keep the email short. Event name, date, one-sentence description, registration link.


In-Person Announcements

If you're on a campus, this is a superpower.

  • Ask professors for 2-3 minutes at the start or end of a lecture to announce the event. Most will say yes if you're respectful of their time.
  • CS classes are obvious, but don't skip business, design, and liberal arts classes. Applied AI is cross-disciplinary. Some of the most engaged attendees at Austin events have been non-technical.
  • Bring a QR code on your phone or a printed flyer so people can register on the spot.

Chapter leaders have proven this model: walked into lectures, gave a quick pitch, and got entire classes to sign up. It works.


Event Platform (Luma, Meetup, Eventbrite)

Your event listing is the landing page, not the promotion channel. People don't browse Luma looking for events. They click through from a social post, a text, or a friend's recommendation.

Make the listing good:

  • Clear title and description (Writing Event Descriptions covers this)
  • Venue address and time
  • Speaker names if confirmed
  • A photo from a past event if you have one

But don't rely on the platform's discovery features to fill your room.


Manufacturing Scarcity (Early Chapters)

Your first few events are the hardest to fill. Nobody knows you, there's no track record, and an empty guest list makes the event look dead. You can use Luma's capacity and visibility settings to create FOMO and tip hesitant visitors into registering.

The progression:

  • 0 registrations: Hide the guest list entirely. An empty list is worse than no list.
  • 10+ registrations: Turn the guest list on. Social proof starts working for you.
  • 20+ registrations: Set the capacity just above your current count so Luma displays “almost at capacity.” Bump it by 5 each time new people register, keeping the event perpetually close to full.

This works because people looking at the page see momentum and a closing window. Both push them to register now instead of later (which usually means never).

Don't overuse it. If your capacity has been “almost full” for three weeks straight, regulars will notice and it starts to feel scammy. This is a cold-start lever, not a permanent tactic. Once you have a few events under your belt, people come for the quality of the room and the impact on their work. Drop the trick when you no longer need it.


UTM Tracking

Every published link to an event must include a utm_source parameter. This lets you see exactly where registrations come from in Luma's Insights tab (or the equivalent analytics on whatever platform you use).

Format: ?utm_source=<platform>-<account>

Examples:

  • https://lu.ma/AppliedAILive004?utm_source=x-chapter-lead
  • https://lu.ma/AppliedAILive004?utm_source=linkedin-chapter-lead
  • https://lu.ma/AppliedAILive004?utm_source=newsletter
  • https://lu.ma/AppliedAILive004?utm_source=x-appliedaisociety

This takes two seconds and gives you real data on which channels are actually driving registrations. Without it, you're guessing. With it, you can see that your LinkedIn posts convert at 3x the rate of X posts, or that the newsletter is your most efficient channel, and adjust your effort accordingly.

Make this a habit. Every link, every time. If someone else is posting on your behalf (a speaker, a partner), give them the link with the UTM already attached.


Partner Amplification

Partners, sponsors, and speakers all have their own networks. Ask them to share the event.

  • Send speakers a pre-written post they can copy and adapt
  • Ask venue partners to include it in their community announcements
  • Tag sponsor organizations in your social posts

See Building Partnerships for more on cultivating these relationships.


What Matters Most

In order of impact:

  1. Personal texts and DMs. Nothing beats a direct, personal invitation.
  2. Consistent social posting with hooks. Multiple times a week, tied to real AI news. Bold all-caps headlines. Event plug on every post (in the reply on X, inline on LinkedIn).
  3. In-person announcements. Especially on campus.
  4. Email lists and partner amplification. Broader reach, lower conversion.
  5. Event platform listing. Necessary but not sufficient. Always UTM-tagged.

The mix will shift as your chapter grows. Early on, it's mostly personal outreach. After a few events, word of mouth and social proof do more of the work.


See Also