We need to talk about AI and Climbing Gyms
Updated: 2 days ago
Much like everywhere else, AI has been a central theme in conversations with clients, potential clients and other folks in the Rise Above network. Is AI good? Is AI bad? How can AI help climbing gyms? Is AI hurting climbing gyms? Etc.
Nobody knows. Including us. Anyone who pretends to know is doing just that, pretending. Typically to sell you their own "AI solutions."
In the last 2 years, since AI has become a more predominant theme at trade shows, in conversations, and as a tool gyms are using more and more, Rise Above has taken a pretty deep dive into how tech platforms, climbing gyms, and other companies are using AI. We have learned about what is working well, and what is failing spectacularly.
So let's talk about it.
NOTE: This article is not about AI companies (of which there are plenty of judgements to be made.) This is about AI as a tool. |
How are Climbing Gyms Using AI?
Up front, I want to state that we have no way of knowing every way every single gym is or isn't using AI. After collecting data on AI usage across 5 countries and dozens of clients and client conversations, here are the main themes, and our rankings out of 5 stars:
AI chat/help bots on gym websites ⭐️⭐️⭐️
Using various "AI Assistant" functionalities to write emails, respond to customers or co-workers, schedule meetings, take notes, etc. ⭐️⭐️
Using AI to screen job candidates 💩
Using LLMs to do reporting and data analytics ⭐️
Using LLMs to help create documents, strategy, tactics and plans 💩
All of these uses seem reasonable on first pass. However, as Rise Above has pulled back the curtain with clients, there is a resounding theme we can not ignore:
Gyms using AI the most, are unanimously performing the worst. Culturally AND from a revenue perspective.
So what does this mean and what might be reasons things aren't exactly going to plan? The Rise Above hypothesis is twofold:
Gyms operators and leaders separating themselves and their people more and more from the day-to-day of the business under the pretense that they are being more "efficient" by having robots do some of that work are losing sight of the important and nuanced realities of a complicated business such as a climbing gym
Gym operators and leaders increasingly leveraging an relying on LLMs to backfill for expertise they don't have in their organization are producing poorer strategy at a linear rate with their use of LLMs.
What are the Risks of AI Use?
It's worthwhile to note what is happening elsewhere right now: companies actually rolling back AI use, even straight up turning it off or seriously guard-railing how teams use it.
80% of AI projects fail to deliver real business value. Only 21% of S&P 500 companies can point to a measurable AI benefit, despite $527 billion in AI spending expected this year.
To be clear, Rise Above is not an AI naysayer. We are not adamantly anti-AI. Rise Above has always been, and always will be rigorously in favor of operational excellence, and that is what informs our probing of this new technology and our desire to identify its highest and best uses.
So far, we have seen more risks than benefits in our industry.
If you have deep subject matter expertise, leveraging LLMs and other AI tools to enhance or speed up your workflows is absolutely the right use. If you don't have marketing expertise (for example), asking an LLM to make you your marketing plan is 80% likely to fail (according to studies by Gartner)
How can Climbing Gyms Navigate AI?
This is for sure the literal million dollar question right now in every industry. In thinking through this for Rise Above clients (and frankly, for our own operations) whilst simultaneously recognizing we are not experts we have done a lot of research. We have landed on two concepts we are successfully implementing with clients today:
Concept 1: Don't let Robots Be in Charge of Your High Risk or High Ambiguity Initiatives & Tasks
My favorite decision matrix came to me recently through a colleague. It's from an amazing article published by the MIT Sloan School of Business.
The concept is simple, think about your AI use through two axis:
Ambiguity (high - low)
Risk (high - low)
The more ambiguous something is, or the higher risk, the more you should consider having human oversight.
We have to remember that AI is a highly advanced prediction engine that takes more data than we can comprehend, passes it through machine learning algorithms and gives you a PREDICTION of what it thinks the right answer is. If you ask AI to write a song (please don't), it will "predict" what a song "should" sound like based on what it has access to, full stop.
Creating music by predicting what a song should sound like feels like the worst possible way to create art in our opinion.
Concept 2: The less Expertise you Have, the More Ambiguous that Topic Is
OK, confession: I am many things, a talented mechanic is not one of them. I'm ok with this. The real problem is that I am not a HORRIBLE mechanic either. What does this mean? It means whenever I need to fix or upgrade something on my 4Runner, it takes me 3-4 times longer than I think it will and I screw it up first go probably 90% of the time. This is just a small example of a well-known concept called the Dunning-Kruger effect. And it is a great metaphor to how we are seeing AI used today.
Don't use AI to Fake It
I would be in big trouble if I asked an LLM to write me instructions for putting a new engine in my truck. Why?
Because I only know a little about this (meaning it's highly ambiguous to me)
I probably know just enough to be over-confident
An engine swap is high risk/impact to the operations of my truck.
If I write a good, detailed prompt on how to do this, I will likely get something back that LOOKS great. The problem?...
I don't know enough about swapping an engine to be able to identify the errors AI has inevitably made so I either waste a bunch of time identifying them late, and/or end up with a truck that doesn't work.
This is exactly what we are seeing over and over in our industry: individuals and teams expecting that an LLM can build their marketing or sales or HR plan. And it is a big problem.
How Can Climbing Gyms Use AI?
If we agree to the risks, the questions becomes: How do we leverage what is an objectively powerful tool to help us operate our businesses?
We have to continue to hire experts, to retain expertise, and to strive to be experts ourselves in the areas that are critical to our business. If we don't have the expertise, we have to backfill the expertise with real actual human-earned expertise.
We have to carefully navigate the decision matrix for how we are using AI based on both ambiguity of the task, and criticality of the outcomes.
We have to accelerate that expertise by using AI tools that make routine activity highly automated so the real expertise can be focused on the high acuity activities.
CONSIDER USING AI FOR THIS
Routine customer service questions and tasks with close oversight from a human with real demonstrated expertise in customer experience. Things like: bookings, membership requests, FAQs, etc.
Collating, aggregating and visualizing data. Always double-checked by a human who knows what they are looking at.
Automating routine tasks such as schedule building, email confirmations, sending out hiring paperwork, etc. Workflows should be built by and iterated by an experienced human.
Research. AI makes mistakes so always ask for sources to be cited.
TLDR: Use AI for routine, low-impact tasks to unleash your experts to focus on high value items and create efficiency and time-savings in areas where little damage can be done. As you learn more, you can assign more to AI.
BE CAREFUL USING AI FOR THIS
Creating strategy and business plans, especially if you are tasking AI because yo don't feel like you yourself know what to do.
Making decisions or recommendations on critical path or highly ambiguous things like hiring or HR matters
Interfacing directly with any stakeholders (customers, members, staff, investors, co-workers, etc) without a high degree of talented oversight.
Almost any version of final creative, design, or artwork. This is a tiny bit of our personal opinion coming in, but still...avoid it. Leave the creation of art to humans. Let the robots do your math.
TLDR: Be very cautious with the implementation of AI when you don't have the real ability to check it against demonstrated expertise and experience.
We don't think the robots are coming for our jobs any time soon.
The reality is, AI is here, it is not going anywhere, and it is an objectively powerful tool with real implications both on the positive and negative side. So we have to be very thoughtful and deliberate in how we build our AI strategies. Agree? Disagree? Want to to talk about what we have seen work and how you can implement AI in your climbing gym? Let's talk.
Comments