Strava
personalized natural-language training insights from activity data
Strava launched Athlete Intelligence in October 2024 to translate training data into personalized natural-language insights: the feature covers pace, heart rate, power, and Relative Effort over a rolling 30 days, in 14 languages, on a base of more than 125 million athletes and 10 billion uploaded activities.
Objective
Strengthen the value of the Strava subscription by translating complex training data into simple, personalized insights, to retain subscribers and make their metrics actionable.
The deployment
Athlete Intelligence is the AI feature launched by Strava on October 3, 2024, first in public beta for subscribers. It analyzes and interprets training data into personalized insights and natural-language guidance. The feature aggregates the trends of the last 30 days of sessions and covers pace, heart rate, elevation, power, and Relative Effort, Strava's proprietary intensity metric. It detects milestones and surfaces highlights: fastest pace, longest distance, highest Relative Effort, greatest elevation. It is available in 14 languages, with an opt-out option. Strava counts more than 125 million athletes in more than 190 countries and more than 10 billion uploaded activities, which gives the system a base of training data at large scale.
Results Proof C
Brand press release (picked up by PRNewswire) plus established tech press naming Strava and the Athlete Intelligence feature. Platform scale figures documented, but no isolated retention KPI or business result, hence C.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- plateforme Couche IA Strava (Athlete Intelligence) in-house generation of natural-language insights and guidance from sessions, opt-in for subscribers
- llm Grand modele de langage translation of the aggregated metrics into personalized text in 14 languages; the provider is not named in the record's sources
- infra Pipeline d'agregation 30 jours rolling aggregation of sessions: pace, heart rate, power, elevation, and Relative Effort, with record detection
How it runs, concretely
For ops teams-
1Activity upload athlete
The athlete syncs their watch or phone; the activity feeds Strava.
-
2Trend aggregation AI / data team
The system aggregates the last 30 days of sessions across the key metrics.
-
3Insight generation AI
An LLM translates the data into a personalized insight and natural-language guidance, in the subscriber's language.
-
4Delivery site_app
The insight and the highlights (pace records, elevation) appear after the activity.
The uploaded activity data (pace, heart rate, power, Relative Effort). Without a sensor or regular uploads, the 30-day aggregation empties out and the insights become generic.
How your customers perceive this type of use
Sourced studiesLe paradoxe est documente des deux cotes : 71% des consommateurs attendent des interactions personnalisees et 76% sont frustres quand elles manquent (McKinsey, 2021), mais 75% declarent ne pas acheter aupres d'organisations auxquelles ils ne confient pas leurs donnees (Cisco, 2024). La « creepy line » est localisee : messages recus quelques secondes apres une recherche et suivi de localisation sont les pratiques qui mettent le plus mal a l'aise (Periscope by McKinsey, 2019).
Acceptance conditions
- La confiance dans le traitement des donnees precede l'achat : 75% ne achetent pas sans elle (Cisco 2024)
- Un cadre legal protecteur rassure : 59% des consommateurs disent que des lois fortes sur la vie privee les rendent plus a l'aise pour partager des informations dans des applications IA (Cisco 2024)
- La personnalisation elle-meme est attendue quand elle est consentie : environ la moitie des consommateurs (US 55%, UK 52%) disent s'inscrire souvent ou parfois a des services personnalises (Periscope by McKinsey 2019)
Red lines
- Le message declenche quelques secondes apres une recherche ou un achat : deuxieme ou troisieme cause de malaise selon les pays (Periscope by McKinsey 2019)
- Le suivi de localisation percu comme de la surveillance : 40% de malaise en Allemagne et au Royaume-Uni (Periscope by McKinsey 2019)
- Le mesusage des donnees personnelles par l'IA, devenu la premiere inquietude des consommateurs, a 53% et en hausse (Qualtrics 2025)
Sources: McKinsey & Company 2021 · Periscope by McKinsey 2019 · Cisco 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- per-user activity or usage feed
- structured performance metrics
- rolling history over several weeks
Org prerequisites
- product team to frame the insight format
- legal basis for data close to health data
- multilingual layer if the audience is international
Possible stack
- managed or in-house LLM
- metric aggregation pipeline
- app with post-activity delivery
The plan, step by step
- Step 1Define 3 to 5 high-value insights (records, 30-day trends) from the metrics already collectedDeliverable: Insight spec with formulas and thresholds
- Step 2Set up the rolling 30-day aggregation pipeline of per-user metricsDeliverable: Aggregates computed and tested on real data
- Step 3Connect the LLM to the aggregates with templates and guardrails (no medical advice), opt-inDeliverable: Post-activity summary generated, in beta
- Step 4Frame compliance: legal basis for data close to health data, user information, opt-out optionDeliverable: GDPR file and functional opt-out
- Step 5Open in beta, measure engagement and perceived value, then extend languages and covered metricsDeliverable: Feature in production, multilingual, with a usage dashboard
First step: Define 3 to 5 high-value insights (records, trends) and connect an LLM to the metrics already available, opt-in.
The same pattern, proven in other industries
Boots
machine learning scoring of offer-usage propensity, applied to mail and the app
Meta
AI feed and video recommendation (large-scale models / LLMs)
Michaels
genAI personalization of marketing messaging (copy) at scale, tuned to the brand voice
Sources
- S1 Strava's Athlete Intelligence Translates Workout Data into Simple and Personalized Insights Interested party archive pending
- S2 Strava's powerful AI insights are here - Athlete Intelligence is now available in beta Established press archive pending
- S3 Athlete Intelligence launches for Strava subscribers Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.