Our Work

EMD UK.

We partnered with EMD UK, the national governing body for group exercise, through our Kickstart discovery and scoping process to explore how artificial intelligence could help fitness professionals spend less time on administration and more time supporting their communities.

EMD UK

Client

Fitness & Wellbeing

Sector

Web app

Product

2026

Year

EMD UK Dashboard

IDENTIFYING THE PROBLEM

Giving Fitness Professionals More Time to Focus on What Matters.

Group exercise plays an important role in supporting physical health, mental wellbeing and community connection. According to EMD UK, 7.2 million adults in England participated in fitness classes during a four-week period in 2024/25. Yet behind every successful class is a considerable amount of work outside the studio, with instructors often responsible for marketing, session planning, participant communications and the administrative demands of running their own businesses.



EMD UK's 2026 Instructor Survey highlighted the pressures facing this workforce. Analysis of 874 active instructors identified five distinct personas, each with different working patterns, professional priorities and support needs. Full-Time Professionals delivered approximately 18 classes per week, reaching an average of 293 participants, while 78% identified financial pressure as a challenge. Community Regulars demonstrated stronger wellbeing but continued to experience pressures around time, finances and participant retention. Meanwhile, Digital Instructors highlighted opportunities for practical business support, class promotion and professional development.



These findings reinforced the importance of developing solutions that respond to the realities of different instructors rather than treating them as a single audience. Artificial intelligence presented an opportunity to reduce administrative pressure, but general-purpose tools lacked the specialist context and verified guidance needed for reliable professional support.



EMD UK therefore recognised an opportunity to explore a trusted, sector-specific AI assistant that could simplify everyday administration while making professional knowledge more accessible.



Key challenges included:

  • Reducing time spent on marketing, class planning and professional documentation.
  • Supporting instructors with different working patterns and levels of digital confidence.
  • Making reliable professional guidance easier to access and understand.
  • Exploring AI-generated content grounded in credible, approved industry sources.
  • Reducing repetitive data entry through personalised profiles and structured workflows.
  • Establishing technical and commercial feasibility before committing to development.
EMD dashboard on a laptop
EDM Web App Tips
EDM Web App
EMD dashboard mobile
AI assistant button
Chair based activities with gentle mobility exercises
5Distinct instructor personas identified
7.2MAdults participated in fitness classes in England within a four-week period
56%Instructors want more training and development
CRAFTING THE SOLUTION

From an Ambitious Idea to a Defined Digital Product.

EMD UK engaged Arch through our Kickstart process to investigate the opportunity and establish a clear development direction. Rather than progressing immediately into a software build, we brought together stakeholder workshops, user journey mapping, interface design and technical assessment to understand what the platform needed to achieve, how it could work and where the principal risks and dependencies lay.



EMD UK's instructor research provided an important foundation for this work. The five identified personas, Digital Instructors, Community Regulars, Peak-Time Side Instructors, Weekend Warriors and Full-Time Professionals, demonstrated the diversity of the workforce and helped inform a task-based experience tailored to different professional needs.



Designing a Practical AI Assistant

Through the scoping process, we established a proposed product centred on three guided AI generators covering marketing content, class planning and professional policies. Each would use structured questions to gather relevant information, allowing instructors to produce useful, editable content without needing to understand complex AI prompting.



Alongside these tools, the proposed platform includes a conversational AI assistant for open-ended questions, content refinement and additional support. A task-first dashboard would keep core functionality accessible, while personalised instructor profiles could store relevant business information and professional context for reuse across the experience, reducing repetitive data entry and improving consistency.

The intention is to make AI a practical extension of an instructor's working routine, rather than introducing another complicated system to manage.



Exploring Trusted, Evidence-Based AI

An important part of the technical assessment was establishing how professional guidance could be supported by reliable sources. We explored a Retrieval-Augmented Generation (RAG) architecture that would connect an established AI model to a curated repository of EMD UK-approved and relevant third-party knowledge.



This approach could enable the platform to retrieve supporting information when answering professional questions, display relevant sources and introduce safeguards where guidance is missing, contradictory or outside the approved knowledge base. It would allow EMD UK to maintain greater control over the professional information available through the proposed assistant while distinguishing verified guidance from general AI-generated content.



Defining the Technology and MVP Roadmap

Our technical review considered the architecture required to support the platform, including a responsive web application, secure accounts, AI orchestration, structured data storage, document management and administrative controls. We also explored potential connections with EMD UK's membership and learning systems, including Sport:80, identifying areas requiring further validation.



By reviewing these considerations before development, we helped shape a proposed MVP that balances functionality, usability and technical feasibility. The resulting scope establishes a phased approach, with the flexibility to introduce additional integrations, commercial access models and AI functionality as the product evolves.

EMD UK Assistant 1EMD UK Assistant 2EMD UK Assistant 3EMD UK Assistant 4EMD UK Assistant 5

THE OUTCOME

A Clearer Route From Concept to Development.

The outcome of our Kickstart engagement was a defined product direction, technical approach and proposed MVP scope for an AI-powered assistant tailored to group exercise instructors.



Through collaborative discovery, user-experience planning and technical assessment, we helped EMD UK explore the opportunity in practical terms, identify important development considerations and establish a clearer route towards implementation.



The Kickstart deliverables covered:

  • Client workshop: Exploration of user needs, business objectives and product priorities.
  • Technical feasibility and architecture: Assessment of the proposed technical approach and infrastructure.
  • Key screens and user journeys: UX planning and interface designs illustrating the core experience.
  • AI, API and integration review: Evaluation of relevant AI technologies, external services and existing systems.
  • Risks, dependencies and limitations: Identification of factors requiring further consideration or validation.
  • Technology recommendations: Guidance on appropriate development frameworks and services.
  • Defined MVP scope and roadmap: A structured development plan covering the initial product and future opportunities.
  • Fixed costs: A clearer commercial basis for the proposed development phase.



Delivered through a focused process typically spanning 2–4 weeks, Kickstart helps organisations move from early-stage ideas to informed development decisions without committing prematurely to a full software build.



For EMD UK, the engagement established the foundations for a platform intended to reduce administrative pressure, make specialist guidance more accessible and strengthen the organisation's digital support for fitness professionals. While development has not yet commenced, the proposed solution is now supported by a clearer product strategy, technical direction and delivery scope.

PROPOSED MVP FEATURES

An AI Assistant Designed Around Everyday Instructor Tasks.

The product scope established through Kickstart identifies the following proposed capabilities:



  • Responsive web application for mobile, tablet and desktop.
  • Secure user accounts and personalised instructor profiles.
  • Task-first dashboard with access to core AI functionality.
  • Guided marketing, class planning and policy generators.
  • Conversational AI assistant for additional support and content refinement.
  • Reusable instructor and business information.
  • RAG architecture incorporating approved professional knowledge.
  • Source referencing and safeguards for professional guidance.
  • Document library with editing, duplication and export functions.
  • Administrative tools for managing knowledge sources and templates.
  • Usage analytics and monitoring.
  • Scalable architecture supporting future integrations and commercial development.


These features represent the proposed MVP scope, rather than functionality that has already been developed.