Event Pros Want to Add AI? Ask These 4 Questions First

I was chatting with a friend the other day.

Her name is Cathy McPhillips. She's the Chief Marketing Officer at SmarterX, and she also oversees several brands under the Marketing AI Institute, including MAICON — an AI marketing conference.

Right in the middle of our conversation, she paused and said: "Hold on, I've got an agent running in the background."

She had tasked an AI agent with screening 100 of the most active practitioners in marketing AI, prioritizing them for an upcoming flash campaign. The agent was just quietly chugging away while she chatted with me, not missing a beat.

"It looks really simple," she said. "But it's getting the work done while I'm doing other things."

I thought to myself: This is exactly how it should work.

The best AI workflows save every spare minute — and hand that time back to the things only humans can do.

What do I mean?

Let's take it step by step.

Before the Event: AI Does the Homework, Humans Make the Calls

At MAICON, the first task Cathy's team handed off to AI was the kind of work nobody wants to do.

What kind of work?

Scraping competitor websites. Taking screenshots. Filling spreadsheets with sponsor packages, logos, pricing. Who wants to do that? Nobody. But AI doesn't mind. It doesn't just pull the data for you — it also lines up who you should call first.

Sometimes it even handles pricing analysis. Occasionally, you look at the numbers and realize a price bump makes sense. Other times, it tells you the opposite: don't raise prices.

Then there's a harder one: forecasting.

Every year at MAICON is different. You had 800 attendees last year — how many will come this year? Just looking at the numbers doesn't tell you much, because each year is a one-off event. But now Cathy can ask her AI the way she'd ask a colleague: "Is this a trend, an anomaly, or did we just get the last forecast wrong?"

Think about that. A colleague who never forgets anything, available to ask and answer anytime.

But Cathy didn't hand over the judgment calls.

The AI helped her recover the institutional memory that judgment depends on — who came last year, who sponsored, where things went wrong. Once she had that context, the call was still hers to make.

And the time saved?

"All those things I used to have to do manually — now I can spend that time being a human. Talking to people. Maintaining the relationships that genuinely need a human touch."

AI didn't replace judgment. It gave judgment the ammunition it needed.

But There's One Thing She Won't Touch

Email.

Email is MAICON's most important channel, and Cathy has drawn a very clear line in the sand.

"People chose to let you into their inbox," she said. "I'm not sure I'm ready for customers to receive something that doesn't sound like us. I want to be their best steward."

It's not that she didn't try.

The team built a persona GPT and was fairly confident in it. The GPT rewrote a batch of emails, and on the surface, they looked reasonable. And then?

Open rates dropped. Click rates dropped.

Cathy turned to AI to diagnose what went wrong. AI told her: these emails no longer answer the customer's questions.

Her inner monologue at that moment was essentially: "You're the one who told me to do this!"

Good grief.

But that experience taught her a lesson that runs through every workflow she builds: AI is not a shortcut around strategy. Whatever you do, you do it with purpose and intention.

The Agenda: Designing the Experience Through the Customer's Voice

MAICON's agenda team did something really smart.

They took a draft agenda and held it up against the voice of their customer. Where does that voice come from? Evaluation forms, Slack discussions, podcast comments, webinar chat logs.

The comparison revealed a gap: attendees who were just getting started with AI felt the content was sufficient. But the veterans — people who'd been using AI for a while — felt the content was too shallow. They'd outgrown it.

So what do you do?

Cathy's team added two new formats.

One is called the transformation stage. Fifteen minutes, one-on-one conversation with a CMO, focused on a single question: what did you actually implement?

The other is the Build session. You show up, open your laptop, and thirty minutes later, you walk out with a working agent.

See, AI helped them see the problem. But the solution was designed by humans.

And then there's one problem that didn't need AI at all.

Every year, chasing speakers for titles and abstracts is a headache for everyone. Cathy's team didn't build a tool — they changed the process: speakers submit their title and abstract right when they sign the agreement. No abstract, no contract.

Sometimes the most effective automation is not automating.

During the Event: A One-Hour Content Machine

Once the event kicks off, things move fast.

Goldcast hooks into the main stage recording. Within an hour of each session ending, ten short video clips are ready to go. Straight to social media.

The next morning, an attendee who missed day one came over and said: "I wasn't here yesterday — I want to catch up." They pulled out their phone, and everything was right there.

By the time the whole event wrapped, they'd accumulated around 150 pieces of content. Next year's teaser campaign? Done.

Even better: the speaker kits.

The day after the event ends, every speaker receives a kit — highlights from their session, questions from the audience, edited short videos, custom graphics. One purpose: you showed up for us, so we'll help your content go further.

"We're working hard to create an experience for speakers where they feel: I'd come back and collaborate with you anytime, because you're helping me too."

How are these kits made? Assembled in a Claude project. But every single item gets reviewed by a human before it goes out.

"We have humans at every step."

That's powerful. I completely agree with that judgment.

After the Event: Finding the Threads in Three Days of Content

What most event teams do after a conference is toss a few recordings onto LinkedIn and YouTube and call it a day.

Cathy thinks that's the biggest waste.

"It's such a shame — there's so much content."

AI helped her team go through all the sessions from the three days and surface recurring themes. Some of those themes were topics that speakers who had never shared a stage had each addressed from different angles. You'd never catch those cross-connections by having a human watch three days of recordings.

And once those threads are found? They become podcasts. Social content. Slack discussions. Webinars.

They're even mapping that content to the questions people are searching for, to optimize for traditional SEO and answer engines.

This is still a work in progress. Cathy is refreshingly honest about it: "We're experimenting, just like everyone else."

The 4 Questions

By now, you might be thinking: I want to bring some AI into my own event workflows. Where do I start?

Cathy offers four questions. Simple, but sharp.

First: Is this lightening your load, or replacing your relationships?

Cleaning up a contact list — let AI handle that, and use the saved time to make phone calls. But if you let AI auto-send those messages too, your relationships start thinning out. Cathy's agent filled in contact details for her, but every single message was sent by her personally. Of those 100 people, she personally knows sixty or seventy of them. Every one of them got a personal note from her.

Second: Does a human review the output before it reaches the customer?

Mistakes damage trust. Speaker kits go through Claude, but every item gets human verification. Because if a speaker receives incorrect feedback, they won't thank you for the speed.

Third: Does this workflow expose customer data to the tool?

Set the rules before you experiment. SmarterX strips out all identity information before feeding any customer feedback into their voice-of-customer knowledge base.

"We put everything in," Cathy said. "We'd even put in our P&L (profit and loss statement). But customer data? No."

Fourth: Can the tool you're already paying for get you 80% of the way there?

SmarterX spent this year's budget on hiring, not on buying more software. Before any new tool walks through the door, it has to answer a few questions: Does the team actually need it? Can it be trusted? Is it compatible with HubSpot?

Cathy's standard is even simpler: "Can the tool you're already paying for get you to 80% first?"

---

Look at these four questions, and you'll see they all point to one thing.

AI tools will iterate. They'll get more powerful. The answers to these questions will change. But one thing won't change: your relationships with customers, speakers, and sponsors are yours.

The trust that needs a human to maintain it. The awkwardness that needs a genuine, heartfelt word to smooth over. These are the places where humans will always add the most value.

That afternoon when I was chatting with Cathy, her agent finished running in the background. 100 contacts, prioritized and ready. But what she did next was sit down and write emails, one person at a time.

Because those emails had to be from her.

May you find your own 80%.