
Spark demonstrations during edUcon helped participants familiarize themselves with the AI tool. (Whatever Media Group)
A lot has changed in the conversation about generative AI since last January, when Anh Nguyen, CMP, founder of Spark Event Management, and Shawn Cheng, co-founder of #EventProfsBreakShit, organized a face-off at Convening Leaders between humans working with AI on event-related challenges and humans working without it.
Since then, as more event professionals have become familiar with AI tools, talk “has shifted away from whether or not AI is a helpful tool for planners,” Sheng told Convene, “to questions about how to use it better and which tools will help them the most.”
To put that to a test, Nguyen and Cheng organized a second contest, “Duel of AI: Finding the Best AI Tools to Replace ChatGPT,” held at edUcon in June, and which once again gave two competing teams the same challenges to solve in a 45-minute session. This time, a team from Gevme, the Singapore-based event tech company which developed Spark AI with PCMA — Jonathan Easton, vice president of design, and Keerat Singh, head of revenue growth — could only use the enterprise version of Spark AI. The second team, made up of two planners, Nyguen and Tess Vismale, CMP, DES, founder and CEO of iSocialExecution, and Vinnu DeShetty, founder and CEO of the Event Tech Hub, could use any AI tool except Spark AI. The audience members served as judges, Cheng said, in what was a loose, rather than rigorous, evaluation process.
It was close race. Here are the challenges the teams faced in the competition’s three rounds, and their outcomes:
Round One
Teams were asked to create three attendee personas — representing an event strategist, a business event supplier, and a student — and a strategy that would convince each to register for an upcoming meeting. The planners, who spent more time upfront crafting prompts for ChatGPT compared with the Spark AI team, were deemed the winners by the audience, although in his view, the results arguably were a tie, Cheng said. “Spark AI came up with a very good, very detailed plan, but it just felt a little cold,” he said, while the planners’ results were “more human-centric.”
Round Two
Each team received a conference session recording and was asked to repurpose the content. The planners’ team turned the content into a series of blog and social-media posts; the Spark AI team produced a six-segment online course. The course was impressive, Cheng said, but the audience gave the planners the win because they felt their content was better suited to the needs of end users.
Round Three
The teams were given registration data from a past conference and asked to use it to create a strategy for improving future attendee registration. In this round, Spark AI was the clear victor, Cheng said. Spark AI is designed to generate that kind of detailed strategic plan in just a few steps, while the planners, Cheng said, wrote prompts for several tools to get a result. In retrospect, Cheng said, there was a flaw in the competition’s design. “If I had put planners on the team that was using Spark AI, I’m wondering if the result would be different,” he said. For him, the biggest takeaway was how often the quality of the results came down to the human who is using the product — in this case, planner or tech developer — and the depth of their knowledge and experience in terms of their audience’s needs.
As AI tools continue to evolve, “I think a lot of planners are like me,” Cheng said. “We’re confident in using AI — and we’re not afraid that it will replace our abilities.”
On the Web
For information on how to access free and for-fee Spark tools and online AI education, go to pcma.org/spark.

Shawn Cheng
AI vs. Diversity?
As the two-year anniversary of the launch of OpenAI’s ChatGPT in November 2022 approaches, event professional and tech expert Shawn Cheng sees event professionals moving into a second phase, he told Convene. The first was sheer amazement at generative AI’s ability to draw on vast amounts of information to create content virtually instantaneously. But some of the excitement has been tamped down on the content creation side, he noted, by studies that have shown it can blunt creativity, on a group level, by reducing diversity of thought.
The second phase, Cheng said, “is diving into AI as an analytical and consulting strategy tool. I think we should actually be more cautious that we learn from the experience we have had with creativity. Because what we also do not want in six months from now, is that the conversation we are going to have again is all those analytical reports sound the same.”
“There is indeed a risk for a reduction in diversity of thought” when using AI, Gevme’s Jonathan Easton told Convene, “and it will probably still happen to some extent. But that can be counterbalanced by us humans adding our own personal and unique perspectives and experiences as much as possible when triggering generative AI to create something. So, the fear that AI will make us all think alike is real — but it underestimates the inherent diversity of human thoughts and experiences and our ability to adapt and innovate.”
AI Users, by the Numbers
According to “The AI Proficiency Report,” released by Section, a New York–based online business school, only 7 percent of knowledge workers in the U.S., UK, and Canada use AI every day — a group that Section termed the “AI Class.” Of that group, the majority report saving more than 10 percent of their time each week as a result, and a third of the daily users report saving more than 30 percent of their time each week. Almost all of those in the AI Class — 99 percent — only use a paid version of an LLM (large language model), and most use it to support three or more kinds of work, such as to save time on repetitive tasks, to create content, as a researcher, and as a thought partner.
The largest group of users — 57 percent — are AI newcomers, people who used a gen AI tool a few times a month, “but haven’t unlocked any real productivity gains,” the report said. A quarter of users are AI experimenters, who use AI once a week and report small productivity gains and the smallest group — 11 percent — are AI skeptics, who almost never use AI.
Barbara Palmer is deputy editor at Convene.