The Intelligence Layer

AI that runs
through every layer.
Not bolted on top.

Most fitness platforms added AI as a feature. Dizios was built AI-first. Here's how that actually works.

Trainer Assignment

How Dizios matches members to trainers.

When a new member joins, Dizios's assignment engine scores every available trainer at that studio against the member's profile.

Five factors weighted:

  • Availability overlap (30%)
  • Specialization match (25%)
  • Current trainer load (15%)
  • Dizios Score (20%)
  • Experience match (10%)

Hard filters apply for gender preferences, language needs, and special cases like pregnancy or U-18 members.

The owner reviews the suggestion with full reasoning, approves or overrides in one tap, and the member meets their assigned trainer.

Session Generation

Every workout, generated and scored.

Sessions are generated using a combination of evidence-based templates and member-specific personalization:

  • Member's goal (weight loss, muscle gain, etc.)
  • Experience level and progression history
  • Equipment available at the studio
  • Banned exercises based on injuries
  • Last session's actual performance and RPE

The output is a complete session with exercise selection, sets, reps, rest periods, load guidance, and coach notes.

Every session is scored on three dimensions before it reaches the member.

The Three Scores

Quality, made measurable.

Goal Alignment Score (0-100)

Measures how well the session's exercise selection matches the member's stated goal. Compares ideal exercise distribution for the goal vs. actual. Each goal has a defined ideal composition.

Safety Score (0-100)

Evaluates risk given the member's profile. Deductions for banned exercises, contraindications, excessive load progression, age-inappropriate intensity, pregnancy considerations, and more. Sessions below 50 cannot be approved.

Progression Score (0-100)

Measures appropriate progression from previous sessions. Tracks load and volume progression at the right pace. Flags excessive jumps or stagnation.

Retention Intelligence

Churn prediction before it happens.

Every member gets a daily health score based on attendance patterns, workout frequency, RPE trends, body metric progression, payment history, and engagement signals.

Members are tagged into cohorts: Committed, Stable, Drifting, At-Risk, Inactive. When a member moves to a worse cohort, the studio gets alerted with a recommended intervention.

This isn't analytics dashboards. It's actionable intelligence.

The Owner's Brain

Your studio, summarized every morning.

At 7 AM every day, Dizios sends the owner a briefing on WhatsApp:

“Yesterday: 87 check-ins (down 8% week-on-week). 3 members at high churn risk: Anjali, Ravi, Priya. Payment failures: 2, both retried. New leads overnight: 4 from Instagram. Trainer Suresh called in sick, Dizios reassigned his 7 AM batch to Karan. One thing to do today: call Anjali, she hasn't shown up in 11 days.”

This is your studio briefing you on itself.

Member-Facing AI

An AI coach for every member.

Members can ask Dizios anything. Form questions, nutrition, supplements, training advice, motivation. All answered by AI personalized to their goal and profile.

Critically, the member-facing AI never validates unsafe advice. If a trainer ever recommends a banned substance or unsafe practice, the AI flags it immediately and escalates to the studio owner.

This is fitness safety, built into every member conversation.

Cost and Performance

Fast. Affordable. At scale.

Dizios's AI runs on Anthropic's Claude models. Routine generation uses Claude Haiku for speed and cost efficiency. Safety-critical conversations use Claude Sonnet for nuanced judgment.

Most session generations resolve in under three seconds. Member AI responses arrive in under five.

We've designed the system to keep AI costs sustainable even as the network scales to thousands of studios.

Want to see the intelligence layer in action?

Book a Demo →