AI Use Case Discovery Sprint for a Premium Fitness Club Group
Let's Connect
Overview
What we built
A premium fitness club group with 12 locations was fielding weekly AI pitches and had no evidence to judge any of them. A three-week discovery sprint turned the noise into 4 pilot-ready business cases.
In plain terms: every week another vendor arrived promising the group an AI breakthrough, from predicting which members would quit to virtual coaching and front-desk chatbots. Nobody could say which pitch would actually pay off, because the evidence needed to judge them, the group's own member data, was split across separate booking, billing and access-control systems. A previous chatbot trial had been quietly abandoned after three months, which left club leadership warier of the whole subject rather than clearer about it.
We ran a three-week discovery sprint to replace pitches with evidence. We interviewed stakeholders across five departments, audited how ready the club systems' data really was, and facilitated workshops where club managers and trainers scored ideas on impact and effort. The sprint captured 31 candidate use cases and narrowed them to 4 pilot-ready initiatives, each with a named owner, a costed business case and a measurable pilot plan. The first pilot, churn-risk flagging for renewals, launched within 8 weeks, and dropping two overlapping vendor contracts trimmed projected annual tooling spend by 22%.
The Problem
Vendor noise, no evidence
The group runs premium fitness and wellness clubs across 12 locations, exactly the kind of business AI vendors love to call on. Pitches arrived weekly: churn prediction for renewals, virtual coaching for members, chatbots for the front desk. Each sounded plausible in isolation, but the leadership team had no shared method for judging which would actually pay off, so every pitch triggered a fresh round of internal debate that ended without a decision.
Underneath the indecision sat a data problem. Member information was fragmented across booking, billing and access-control systems, so the group could not see one member's full story: what they booked, what they paid and how often they actually walked through the door. Any of the pitched tools would have had to be judged, and eventually trained, on data the group could not yet assemble in one place.
The group had also been burnt before. A front-desk chatbot trial had been quietly abandoned after three months, with no clear account of why it failed or what a better attempt would need. That history made club managers sceptical of the next initiative, and it meant every new vendor conversation started from zero rather than from lessons already paid for.
Weekly vendor noise
Churn prediction, virtual coaching and front-desk chatbots were pitched week after week, with no shared way to compare their claims or their fit with the clubs.
Fragmented member data
Booking, billing and access-control systems each held part of the member picture, so no tool could be evaluated against a complete view of the member.
No evaluation method
Ideas were debated on instinct and vendor promises rather than scored, so decisions stalled and the same arguments repeated with every new pitch.
A failed trial's shadow
The earlier chatbot trial was quietly abandoned after three months, leaving scepticism behind and no record of what the next attempt should do differently.
What it was costing them
Every stalled decision carried a price. Managers and trainers lost hours to vendor demos that went nowhere, member data stayed too fragmented to support any serious tool, and the abandoned chatbot trial kept feeding scepticism. Meanwhile clubs that could not flag members at risk of leaving were learning about departures only after they happened, with no chance to intervene.
The Solution
Three-week discovery sprint
We designed the engagement as a three-week sprint with a hard finish line: leadership should leave with decisions it could act on immediately, not another study. The first strand was listening. Stakeholder interviews across five departments surfaced where work actually hurt, from renewals to the front desk, and captured the ideas staff had been quietly carrying alongside the ones vendors had planted.
The second strand was evidence. A data-readiness audit of the club systems established what the booking, billing and access-control data could genuinely support today, so no use case could be shortlisted on hope alone. With interviews and audit in hand, we facilitated impact-versus-effort workshops where club managers and trainers scored all 31 candidate use cases together, against the same criteria, in the open.
The sprint closed with a shortlist built for action. The 4 pilot-ready initiatives each left the room with a named owner, a costed business case and a measurable pilot plan, so the leadership team could fund them without another round of analysis. The audit also exposed two overlapping vendor contracts duplicating each other, which the group dropped.
Key decisions
Three weeks, not three months
A fixed sprint window forced choices: every use case either earned its evidence within the three weeks or waited, so momentum never drained into open-ended study.
Listen before shortlisting
Stakeholder interviews across five departments came first, so the shortlist reflected where the clubs actually hurt rather than which vendor had pitched most recently.
Audit data readiness early
The booking, billing and access-control systems were audited up front, so every score carried an honest view of what the data could actually support.
Score with the operators
Club managers and trainers, the people who would live with any tool, ranked the use cases themselves in impact-versus-effort workshops.
Ship business cases, not slideware
Each shortlisted initiative left with a named owner, a costed business case and a measurable pilot plan, ready for an immediate funding decision.
Measurable Impact
What changed after launch
Three weeks of structured work replaced months of circular debate. From 31 candidate use cases captured across 5 departments, the group emerged with 4 pilot-ready initiatives it actually believed in, each costed and owned. The first pilot, churn-risk flagging for renewals, launched within 8 weeks of the sprint, giving clubs a way to spot members at risk before they walked away.
The commercial housekeeping paid off too. Dropping two overlapping vendor contracts trimmed projected annual tooling spend by 22%, money recovered before any new tool was bought. Just as durably, the group now has a standing method: new AI pitches are scored on impact and effort like everything else, so the weekly vendor noise now lands on a process rather than a nerve.
Decision making
Weekly vendor pitches, endless debate, no decisions
Leadership acting on 4 costed, pilot-ready business cases
Member data
Fragmented across booking, billing and access-control systems
Audited, with readiness understood before any tool decision
Churn response
Members lost with no early warning
Churn-risk flagging pilot live within 8 weeks
Tooling spend
Two overlapping vendor contracts running in parallel
Contracts dropped, projected annual tooling spend trimmed 22%
Headline results
31 candidate AI use cases captured and scored across 5 departments in 3 weeks
Shortlist narrowed to 4 pilot-ready initiatives, each with a costed business case
First pilot, churn-risk flagging for renewals, launched within 8 weeks of the sprint
Two overlapping vendor contracts dropped, trimming projected annual tooling spend by 22%
Tech & Tools Used
What powered the build
Every tool below earned its place in this engagement. Here is the part each one played.
Miro
Hosted the impact-versus-effort workshops where club managers and trainers scored the candidate use cases together against one shared set of criteria.
Airtable
Ran the use-case register through the sprint: every captured idea with its scores, owner and shortlist status, visible to all departments.
Notion
Holds the costed business cases and pilot plans handed to leadership, so each shortlisted initiative stays traceable from workshop to funding decision.
Looker Studio
Powered the readout dashboards that showed leadership how the shortlist scored and what the data-readiness audit found at each club.
Python (scikit-learn)
Used to prototype the churn-risk flagging approach on sample member data, testing whether the shortlisted pilot was feasible before it was costed.
Google BigQuery
Held consolidated extracts from the booking, billing and access-control systems, giving the data-readiness audit one place to examine the member picture.
Fivetran
Replicated data out of the club systems into BigQuery without custom pipeline work, so the audit started within days of kick-off.
Metabase
Gave club managers self-serve views of the audited data during workshops, so scoring debates could be settled by looking rather than guessing.
Slack
Kept the sprint moving across 12 locations: interview scheduling, audit questions and workshop follow-ups handled in shared channels.
Ready to Build your Fitness & Wellness Business with AI Use Case Discovery
Ask Byte
Ask Byte
Typically replies instantly
just Now
Hi! I'm OrganByte's assistant. How can I help you today?
AI-generated content may be incorrect
