Guidesly has deployed an AI system called Jack AI on Amazon Web Services to automate trip report generation for outdoor guides. The platform processes trip media through AWS Lambda and AWS Step Functions to structure and route data. Media files land first in Amazon Simple Storage Service (S3) before being analyzed. The system uses Amazon Relational Database Service (RDS) to store guide profiles and trip metadata reliably. Computer vision and generative AI models run on Amazon SageMaker and Amazon Bedrock to extract highlights and write narratives. Final reports are formatted for marketing use and distributed across multiple channels without manual editing.
The project addresses a bottleneck in outdoor guide businesses where producing post-trip content slows down operations. Guidesly’s clients include rafting companies and hiking tour operators who need consistent reports to attract new customers. By automating narrative creation, guides spend less time on paperwork and more time leading trips. Security and scalability were built into the system from the start to handle sensitive trip data and growing user loads.
AWS services were selected to meet three requirements: fast processing of large media files, consistent performance under load, and tight integration between components. Amazon S3 stores raw trip photos and videos while AWS Lambda functions trigger analysis pipelines. AWS Step Functions coordinates the sequence of AI tasks to ensure reports are generated in the correct order. Guidesly’s engineering team reported processing times under three minutes per trip report during initial tests.
The AI models were trained on outdoor activity data to recognize landmarks, wildlife, and weather conditions. Amazon SageMaker handles the computer vision tasks that identify key moments in trip media. Amazon Bedrock runs the language models that convert these moments into written narratives. Guidesly chose these services to avoid managing custom infrastructure while maintaining control over model behavior and output quality.
Company representatives stated the system has reduced report generation time by 80% compared to manual methods. They also noted improved consistency in report style and accuracy in content details. The platform currently supports English and Spanish with plans to add French and German by the end of 2024.
Source: aws.amazon.com