The fitness industry has spent years talking about the power of group exercise in driving retention, building community and helping members form lasting habits. Yet for many people, there remain significant obstacles to walking through the studio door for the first time.

For all the sector’s advances in programming, timetable innovation and boutique launches, the customer journey around group exercise has often remained surprisingly blunt: open the app, scan a timetable full of unfamiliar class names and try to work out where you belong.

“The industry has become very good at giving people choice,” observes Danny Woods, CEO of Connect. “Where we still struggle is in helping people navigate it.”

It is this gap – between offering experiences and actively guiding people towards them – that’s being addressed by a new wave of AI-enabled tools. Their focus: not on replacing instructors, nor automating the highly people-centric studio experience, but on reducing friction, uncertainty and the number of moments where someone quietly decides group exercise (GX) probably isn’t for them after all.

Between them, these emerging technologies address a series of critical moments in the GX participation journey: helping members discover what to do; helping them feel confident enough to do it; helping operators understand what’s actually happening in their studios; and helping businesses reinforce the habits that keep people coming back.

“The industry has become very good at giving people choice, but not at helping them navigate it”

Discovery & recommendations

Let’s start with Connect’s new LeisureWise product – a recommendation engine that integrates with operators’ existing tech stacks, embedding into member-facing apps to personalise the path into activity.

“When somebody first joins a gym, LeisureWise asks a few simple questions around goals, confidence levels, preferred training times and activity interests,” explains Woods. “It then recommends classes, activities and experiences based on those inputs.”

The top layer of recommendations are based directly on the member’s responses, followed by experiences enjoyed by others with similar profiles. The third layer – ‘why not also try’ – is non-targeted, allowing operators to strategically promote activities. For each activity, the member is shown the schedule – linked to a booking option if the MMS allows – as well as a description of the activity, what to expect and what to bring.

Crucially, LeisureWise also explains why something has been recommended. “If we’re using somebody’s data to shape their experience, transparency matters,” says Woods.

“Increasingly, people aren’t searching by operator brand. They’re searching for the experience they want.”

Of course, some individuals will know exactly what class they’re looking for – including among non-members and new prospects. In this era of AI, that has a direct impact on how operators structure and present their class information.

Steffie Bryant, president of Keepme, explains: “Increasingly, people aren’t searching by operator brand. They’re searching for the experience they want. It means operators must now structure their data so platforms such as ChatGPT, Gemini and Perplexity can surface relevant experiences through conversational search.”

Emotional reassurance

But even with clearer discovery, uncertainty doesn’t disappear. The barriers to finding a class might be gone, but for many members and prospects, it’s replaced with another: actually walking into one.

This is where conversational AI is beginning to reshape expectations.

“Our sector has talked about personalised journeys for years, but the reality hasn’t historically lived up to the vision,” notes Tamara Bailey, customer success manager at HireBob. This is now changing – and fast – with the growth of agentic AI tools such as HireBob.

Integrated into operator websites and apps, HireBob responds to any chat or voice enquiry in real time. “Ask it anything, at any point, and it instantly provides an answer that’s personalised, relevant, conversational and empathetic,” explains Bailey.

“If someone says they’re interested in indoor cycling, but it’s their first time, the issue isn’t timetable information. It’s reassurance.”

Unlike traditional chatbots, HireBob’s focus is on intent rather than keywords. “If someone says they’re interested in indoor cycling, but they’re nervous because it’s their first time, the issue isn’t timetable information – it’s reassurance,” Bailey observes.

Understanding and responding to this is critical, with group exercise intimidation a major barrier to participation: fear of not keeping up, not knowing what to expect or just feeling out of place.

Where a traditional chatbot might simply share a timetable link in response to any GX enquiry, HireBob’s agentic AI conducts a needs analysis and responds appropriately. It can explain what to expect, suggest arriving early for bike set-up, reassure nervous users they can sit at the back, even propose a studio tour before booking.

“The conversation can continue for as long as the individual needs support, whether they’re a member or a prospect,” adds Bailey.

“It’s why we use the term ‘AI employee’ rather than ‘AI agent’ – because HireBob operates as an extension of your team. It communicates with customers in a way that feels human and on-brand, without ever getting frustrated or having a bad day.”

She continues: “All our AI employees are trained in four things: the operator’s tone of voice and brand personality; the interventions it should choose from; the knowledge base it can draw on; and the goal(s) it’s working towards, whether that’s selling a membership, improving visit frequency, supporting retention or enhancing customer service.”

Operational intelligence

Crucially, every HireBob interaction feeds in to the system’s ‘Business Optimisation Brain’ to generate operational intelligence. Via their own unique portal, operators can tap this data to identify recurring questions, behavioural patterns and friction points at scale, across thousands of member interactions.

“A gym might discover its members are repeatedly confused about booking windows or frustrated by timetable changes,” says Bailey. “Instead of relying on occasional surveys or anecdotal feedback, operators gain visibility into what large volumes of members are consistently experiencing in real time. This allows them to optimise what’s happening across the full customer journey – all the while providing members with immediate, personalised, relevant support.”

“Our tool instantly predicts the likely occupancy, financial and operational impact of any timetable change”

imin’s new Smart Timetable product continues on this theme. “When we started building this tool, our goal was to help operators genuinely understand what’s happening inside their studios,” says head of partnerships Dominik Nilsson.

He argues that, while operators hold reams of attendance and booking data within their leisure management systems, historically very little of it has been surfaced in a meaningful or actionable way. By contrast, Smart Timetable pulls together occupancy data, historical performance, anonymised competitor data, attendee profiles, class reviews, costs, revenues and instructor availability to help operators understand not only which classes are full, but which are genuinely working.

It then instantly predicts the likely occupancy, financial and operational impact of any timetable change, helping optimise the schedule for operators and members alike.

But at least as important is what happens beneath this operational layer.

Supporting best practice

“Group exercise isn’t just about filling classes,” says Nilsson. “It’s about creating experiences people want to come back to. And so we’ve started looking not just at occupancy, but at loyalty. Which classes are building long-term habits? Which appeal to newer members? Which instructors are creating environments where people feel supported and welcomed? Are any classes particularly popular among guests, acting as a strong conversion tool?

“And then on the flip side, are there any classes that look successful on paper, but that might actually be putting people off?

“In indoor cycling especially, regulars in the front row can create fantastic energy, but can unintentionally make a class feel intimidating for first-timers. And so there’s another question sitting behind the data: how many people try that class once and never return?

“Users can rate recommendations – thumbs-up or thumbs-down – creating a continuous feedback loop”

“Traditional occupancy reporting rarely reveals this, so we’re actively exploring how we can surface it within the imin platform. If operators can identify where this is happening, they can start asking better questions about onboarding, instructor approach and studio culture, implementing changes and delivering team training as needed.

“If we can help operators make smarter decisions, help instructors feel more supported and help members feel more engaged and comfortable in returning for that second or third class, then we think technology is doing its job properly.”

Built around behaviours

In this respect, there is further insight to be gained directly from the member.

“LeisureWise 2.0, launched last month, expands its focus beyond new joiners to also actively embrace existing members,” says Woods. “Users can now rate recommendations directly within the platform – thumbs-up or thumbs-down – creating a continuous feedback loop that helps the engine learn what they do and don’t enjoy.

“This is actionable business intelligence that operators can also use to continually optimise the experience and create better outcomes at scale.”

Better outcomes are also central to the Keepme proposition, says Bryant, as she takes up the story.

“The fitness industry has talked about personalised member journeys for a long time, but most remain highly fragmented,” she observes. “A prospect might tell a club they enjoy group exercise and want to lose weight, for example, but those data points disappear the moment they sign up. Valuable intent data is captured, then effectively lost.”

Keepme’s agentic AI platform – Keepme Antares – steps into this void, its five agents sharing consistent data reference points and a common intelligence layer to keep data alive throughout the member journey. Stitching together sales, retention, member services, inbound voice and answer engines such as ChatGPT, member behaviours are continuously tracked and analysed against original stated goals.

“It ensures a completely contextualised conversation at every touchpoint and provides the sort of continuity that’s been missing from the sector for years,” says Bryant.

If attendance starts to drop or engagement patterns change, the platform can intervene early, she adds: “AI is particularly good at recognising behavioural patterns at scale. We can identify member profiles that are more likely to stay if they attend two group exercise classes a week, for example, or when someone might benefit from trying another class format. Off the back of this, we can recommend timely interventions.

“Instead of generic retention campaigns, operators can start building journeys around actual member behaviours, preferences and goals.”

People still matter

Yet even the most advanced technology can only take you so far. “When it comes to the genuine personalisation of group exercise, instructors are still at the centre of everything,” says Nilsson.

Bryant agrees: “We can leverage AI to influence habits, attendance and engagement. We can help get the right people into the right classes at the right time. But operators still have to decide what kind of experience they want members to have once they get there – and instructors still have to deliver it.”