---
title: "How boAt delivers personalized fitness coaching with Amazon Bedrock"
description: "boAt, a leading Indian consumer electronics brand, launched the AI Coach feature on its smartwatches using Amazon Bedrock and AWS Lambda to deliver personalized, actionable fitness coaching. The system ingests encrypted user metrics and preferences from the boAt Crest app, stored in Amazon RDS and Amazon DocumentDB, using historical context to generate plans via the Meta Llama 3.3 70B Instruct model on Amazon Bedrock. To improve personalization, the team made JSON keys more descriptive and added watch-supported activity lists in prompts, reducing unsupported recommendations. A parsing layer with safe defaults handles missing keys, and configurable nudges in various styles help users adhere to their plans. The serverless architecture, chosen for model-agnostic access, built-in observability, and zero infrastructure overhead, has reliably served millions of users across India for over six months."
---

# How boAt delivers personalized fitness coaching with Amazon Bedrock

[AWS](https://yomu.fyi/company/aws) · Natasha Sah · Oct 29, 2026

**Type:** Problem & solution

## Summary

boAt, a leading Indian consumer electronics brand, launched the AI Coach feature on its smartwatches using Amazon Bedrock and AWS Lambda to deliver personalized, actionable fitness coaching. The system ingests encrypted user metrics and preferences from the boAt Crest app, stored in Amazon RDS and Amazon DocumentDB, using historical context to generate plans via the Meta Llama 3.3 70B Instruct model on Amazon Bedrock. To improve personalization, the team made JSON keys more descriptive and added watch-supported activity lists in prompts, reducing unsupported recommendations. A parsing layer with safe defaults handles missing keys, and configurable nudges in various styles help users adhere to their plans. The serverless architecture, chosen for model-agnostic access, built-in observability, and zero infrastructure overhead, has reliably served millions of users across India for over six months.

## Context

boAt, a leading Indian consumer electronics brand, launched the AI Coach feature to provide personalized fitness coaching to its smartwatch users. The system processes user data and preferences from the boAt Crest app, along with historical data, to generate tailored plans. It uses Amazon Bedrock with Meta Llama 3.3 70B Instruct, AWS Lambda for data transformation, and Amazon RDS/DocumentDB for storage. The architecture emphasizes personalization, affordability, and scalability.

## Approach / What changed

AWS Lambda retrieves historical user data from Amazon RDS and DocumentDB, transforms it into a verbose format for LLM understanding, and invokes the Meta Llama 3.3 70B Instruct model on Amazon Bedrock for plan generation. The prompt includes expanded JSON keys and a list of watch-supported activities to improve personalization. Lambda parses the model output, substitutes safe defaults for missing keys, and returns a consistent response to the companion app. Daily nudges are sent based on actual versus recommended metrics, with configurable communication styles.

## Takeaways

- Making JSON keys more descriptive (e.g., weight into weight\_in\_kg) significantly improved LLM’s ability to interpret user data and generate personalized plans.
- Including a dedicated section in the prompt listing watch-supported activities reduced unsupported recommendations, but due to non-deterministic LLM behavior, a parsing layer with safe defaults was still needed.
- Amazon Bedrock was chosen for its model-agnostic API, built-in CloudWatch observability, and serverless scaling, providing clear per-user unit economics without custom metering.

## FAQ

### How does the system ensure recommended activities match the smartwatch's capabilities?

The team added a dedicated section in the prompt listing watch-supported activities. However, because LLMs are non-deterministic, they also added a parsing layer with safe defaults to catch any remaining mismatches.

### What steps were taken to improve plan personalization from initial tests?

Initially, plans were not personalized enough. The team expanded JSON keys to be more descriptive (e.g., weight to weight\_in\_kg, target to step\_goal) so the LLM could better interpret the data, which significantly improved personalization.

### Why did boAt choose the combination of AWS Lambda and Amazon Bedrock?

They selected it for three reasons: model-agnostic access to evaluate multiple FMs via a single API, built-in observability and cost governance with per-invocation metrics, and serverless scalability with zero infrastructure overhead.

**Tags:** [Amazon Bedrock](https://yomu.fyi/topic/amazon-bedrock), [AWS Lambda](https://yomu.fyi/topic/aws-lambda), [Fitness](https://yomu.fyi/topic/fitness), [LLMs](https://yomu.fyi/topic/llm), [Personalization](https://yomu.fyi/topic/personalization)

- Source: [AWS](https://aws.amazon.com/blogs/architecture/how-boat-delivers-personalized-fitness-coaching-with-amazon-bedrock)
- Source URL: https://aws.amazon.com/blogs/architecture/how-boat-delivers-personalized-fitness-coaching-with-amazon-bedrock
- Ingested by Yomu: 2026-10-29T21:00:11.934Z

[Read original post](https://aws.amazon.com/blogs/architecture/how-boat-delivers-personalized-fitness-coaching-with-amazon-bedrock)
