Impediments to AI matchmaking at events

artificial intelligence matchmaking at events Companies are now marketing services for artificial intelligence matchmaking at events. However, unresolved issues could impede the adoption of this technology, especially by attendees.

Consider this marketing pitch for an artificial intelligence event matchmaking service:

“Using the [AI] platform…it’s easier for attendees to make sure they have the right meetings set up, and for exhibitors to have a higher return on investment in terms of connections with high-quality buyers.”
—Tim Groot, CEO Grip, as quoted in What AI Means To Meetings: How Artificial Intelligence will boost ROI, Michael Shapiro, July 2017 Meetings & Conventions Magazine

A win-win for exhibitors and attendees?

Tim describes using artificial intelligence matchmaking at events as a win for both exhibitors and attendees.

I’m skeptical.

Let’s assume, for the moment, that the technology actually works. If so, I think suppliers will reap most of the touted benefits, quite possibly at the expense of attendees. Here’s why.

Successful matchmaking needs digital data about attendees. An AI platform cannot work without this information. Where will the data come from? Tim explains that his service builds a profile for each attendee. Sources include “LinkedIn, Google, and Facebook”, while also “scouring the web for additional information”.

Using social media platform information, even if attendee approval is requested first, creates a slippery slope, as privacy issues in meeting apps remain largely undiscussed and little considered by attendees during the rush of registration. The end result is that the AI matchmaking platform gains a rich reservoir of data about attendees that, without strong verifiable safeguards, may be sold to third parties or even given to suppliers.

In addition, let’s assume that exhibitors get great information about whom to target. The result: “high-value” attendees will be bombarded with even more meeting requests while attendees who don’t fit the platform’s predictions will be neglected.

In my opinion, the best and most likely to succeed third-party services for meetings are those that provide win-win outcomes for everyone concerned. Unfortunately, it’s common (and often self-serving) to overlook a core question about meeting objectives —whom is your event for? — and end up with a “solution” that benefits one set of stakeholders over another.

How well will artificial intelligence matchmaking at events work for attendees?

Artificial intelligence is hot these days, so it’s inevitable that event companies talk about incorporating it into their products, if only because it’s a sure-fire way to get attention from the meetings industry.
I know something about AI because in the ’80s I was a professor of computer science, and the theory of artificial neural networks — the heart of modern machine learning — was thirty years old. AI had to wait, however, for the introduction of vastly more potent technology to allow practical implementation on today’s computers.

While the combination of powerful computing and well-established AI research is demonstrating incredible progress in areas such as real-time natural language processing and translation, I don’t see why sucking social media and registration data into a database and using AI to look for correlations is going to provide attendee matchmaking that is superior to what can be achieved using participant-driven and participation-rich meeting process combined with attendees’ real-time event experience. (Once again, exhibitors may see a benefit from customized target attendee lists, but I’m looking for win-win here.)

From the attendee point of view

When attendees enter a meeting room there’s a wealth of information available to help make relevant connections. Friends introduce me to people I haven’t yet met. Eavesdropping on conversations opens up more possibilities. Body language and social groupings also provide important potential matchmaking information. An AI matchmaking database includes none of these resources. All of them have led me (and just about everyone who’s ever attended meetings) to professional connections that matter.


I’ll conclude with a story. The June 2017 PCMA Convene article Can Artificial Intelligence Make You a Better Networker? describes a techsytalk session by Howard Givner where he “gave particular emphasis to the importance of facilitated matchmaking at events.” I like to think that Howard discovered this when he attended the participant-driven and participation-rich EventCamp East Coast I designed and facilitated in 2010, about which he wrote:

“…it was one of the most innovative and eye-opening professional experiences I’ve had. Aside from coming back with lots of new tips and ideas, I easily established triple the number of new contacts, and formed stronger relationships with them, than at any other conference I’ve been to.”

We didn’t use an AI matchmaking service.

5 thoughts on “Impediments to AI matchmaking at events

  1. Great post as always Adrian – I will disagree on some parts of it though.

    Before I start, I don’t mean my comment as an endorsement to the provider or to any provider of AI. I am not interested in that game, I actually never post names. Sorry for the disclaimer.

    I believe the point here is not whether an AI app will make a difference. It’s whether an app will make a difference at all.

    An AI app is better than an app. General lists of attendees with low signals for matchmaking are wildly ineffective, they cause the same pains you mention in relation to AI if not worse. I believe trying to better off a basic way of matchmaking through tech, it’s a valid proposition. If AI is the way, welcome AI.

    The real question is (where I 100% agree with you) if apps will solve the flaws of poorly designed events that do not connect humans. You can whack as much tech as you want and load it up with APPs but the effects will always be poor.

    If a meeting is properly designed – and reading this website is a must for those that want to run effective conferences – and an app is well positioned in the tech mix, well communicated, planned with boundaries in mind, then AI is only going to be helpful.

    Limitations like allowing a certain number of meeting requests, identifying key users and letting them decide whether to use the app or not. All great things.

    To make a parallel, do you think that LinkedIn is completely annoying? It probably is but I’ve done great business with it, there are things I can and cannot do, there is an etiquette and if we accept to be there, we acknowledge we use it for business networking. If you are a desirable prey, you probably won’t even sign up.

    Would lLinkedIn benefit from some AI injection? Of course. The current matchmaking based on email SUCKS. I would love to get referrals based on other signals. Facebook is much better at that.

    Going back to our industry though, how is the event is using the app? Are they delegating networking just to the app? WRONG. You will fail. Are they for example creating a space where to use the app? Are they cleverly incorporating the app in some participant driven sessions to get attendees to work together in groups? In that case AI will be helpful.

    1. Julius, first, thank you for your kind remark about the value of reading my blog :).

      We are in agreement about the majority of what you say, namely the place of apps as only one piece of the much larger context of providing effective connections and matchmaking at events.

      My main concern, apart from the privacy considerations that I mention, is that adding “AI” into a matchmaking app may be more of a marketing/buzzword gimmick than something that is truly useful. I’m not even convinced (yet) that it’s a net plus, because once we start offering recommendations on the basis of impenetrable and sometimes non-understandable algorithms, the only test becomes whether we find the results to be more useful than what we were getting before.

      And the jury on that in this area is still out. While it’s becoming clear that realtime AI-driven voice recognition and machine translation and image recognition are successful and useful technologies, I have to say that Facebook’s algorithm that determines what I am shown irritates me no end and, from my POV, is NOT helpful. I’d much rather see everything and filter it myself. Facebook only added their (IMO) crappy algorithm because they discovered that people with thousands of “friends” became overwhelmed by the resulting unfiltered feed and stopped using Facebook.

      Anyway, to reiterate, the important point is that, as you say, apps won’t “solve the flaws of poorly designed events that do not connect humans”. My mission is to advocate for event designs that appropriately optimize connection, and I know you are in favor of that outcome.

      Always great to hear from you Julius!

  2. Are we to assume that social media holds all the information about everybody? And if people are not using social media regularly, do they not count anymore?

    1. Good point David. Yes, these approaches are not going to do so well with attendees who shun social media or who do not give permission for their social media platform accounts to be added to at AI matchmaking database.

Leave a Reply

Your email address will not be published. Required fields are marked *