loader animation

AI-Powered Content Discovery for Digital Health

Reducing manual content management by up to 70% with AI-powered classification, intelligent search, and personalized health recommendations

Scroll down to read more

Project highlights

Industry: Healthcare
Client services: AI Business Transformation
Start date: February 2026
Location: United States
Team size: 6 members 
Duration: Ongoing

About the client

Our client is a healthcare technology company specializing in digital health solutions that promote health literacy, emotional resilience, and well-being among young people.

Its digital platform delivers evidence-based interventions designed to support individuals at risk of mental health crises. Used by families, schools, colleges, and healthcare facilities, the platform provides educational and therapeutic content to help users develop healthier behaviors and strengthen emotional resilience.

Business challenges 

The client’s digital health platform offered an extensive library of evidence-based educational videos and therapeutic content. However, manual content management and limited discovery capabilities made it increasingly difficult to scale operations and deliver personalized user experiences.

Several challenges were limiting the platform’s effectiveness:

  • Time-consuming content classification: Each video and learning module required manual tagging against a complex taxonomy of protective factors, creating significant administrative overhead.
  • Underutilized content library: Valuable information embedded in videos and transcripts was difficult to locate and reuse without intelligent search capabilities.
  • Limited content discoverability: Users lacked intuitive ways to identify relevant modules, potentially reducing engagement and content completion.
  • Strict healthcare safety requirements: AI-generated classifications and recommendations needed to remain transparent, controlled, and subject to human oversight.
  • Fragmented analytics: Reliance on external tracking services restricted visibility into user activity and flexibility in developing future reporting and billing capabilities.

The company needed a scalable AI-powered solution that could streamline content operations, improve user engagement, and support future digital therapeutic capabilities without compromising safety or governance.

Goals set to Achievion

Achievion was engaged to design and develop an intelligent AI layer that would enhance the client’s existing digital health platform with automated content processing, semantic search, personalized recommendations, and advanced analytics.

The primary objectives included:

  • Automate content classification by using AI to identify protective factors and generate suggested tags for human review and approval.
  • Unlock existing content through semantic search that makes videos, transcripts, and relevant clips discoverable using natural-language queries.
  • Improve user engagement by guiding individuals toward relevant learning modules and delivering personalized content recommendations.
  • Maintain AI safety and transparency through approved content, controlled outputs, human oversight, and auditable AI workflows.
  • Enable data-driven decisions by establishing in-house analytics capabilities for monitoring user activity and supporting future reporting requirements.

The goal was to make the platform more intelligent and scalable while ensuring AI enhanced, rather than replaced, professional judgment.

Solution 

Achievion developed an AWS-powered AI layer that connects intelligent content classification, semantic search, personalized recommendations, and user activity analytics within the client’s digital health ecosystem.

At the core of the solution are automated processing pipelines that analyze educational content, extract meaningful information, and organize it into searchable data structures.

Using large language models through Amazon Bedrock, the system identifies relevant protective factors in video transcripts and generates structured summaries and embeddings for intelligent content retrieval.

AI-generated classifications are reviewed and approved by authorized staff before being incorporated into the platform, ensuring that human oversight remains central to sensitive healthcare workflows.

The solution combines several core capabilities:

  • AI-assisted content classification: Analyzes video transcripts, identifies protective factors, and suggests standardized tags to reduce manual processing.
  • Intelligent content processing: Generates module summaries and vector embeddings to support semantic search and recommendation workflows.
  • Natural-language semantic search: Helps staff discover relevant videos, transcripts, and individual content segments without relying on exact keyword matches.
  • Guided content discovery: Makes it easier for users to navigate educational resources and identify modules relevant to their needs.
  • Personalized recommendations: Uses content relevance and user activity to recommend approved learning modules and encourage continued engagement.
  • Human-in-the-loop governance: Keeps AI-generated outputs subject to appropriate review, approval, and traceability.
  • Integrated analytics: Establishes greater control over user activity data, reporting, and future platform optimization.

The architecture integrates with the client’s existing backend and content management systems, allowing AI capabilities to become part of established content creation and delivery workflows.

Key features of the product in detail

AI-Assisted Content Classification

The system automatically analyzes video transcripts to identify protective factors associated with emotional resilience and well-being.

Instead of manually reviewing and tagging every piece of content, administrators receive AI-generated classification suggestions aligned with the platform’s established taxonomy.

Staff can review, adjust, and approve the proposed tags before publication, reducing repetitive work while maintaining consistent content standards and human control.

Intelligent Video Search and Content Reuse

Semantic search helps content creators and administrators discover relevant information across the platform’s existing video library.

Users can search using natural-language queries and identify matching videos, transcript passages, and relevant timestamps without needing to know the exact terminology used in the content.

A dedicated desktop application further supports content exploration and extraction of relevant video clips, making existing educational resources easier to find, reuse, and repurpose.

Personalized Learning Recommendations

An intelligent recommendation engine helps users discover approved educational modules relevant to their interests and previous activity.

By considering watched content and module relevance, the system supports a more personalized learning experience, reducing the effort required to navigate an extensive content library.

Guided discovery and reminders are designed to encourage ongoing participation, improve module completion, and help users engage with relevant health resources.

AI Evaluation and Safety Controls

The solution incorporates evaluation and governance mechanisms designed for the sensitive requirements of digital health applications.

Rather than relying on unrestricted AI conversations or opaque recommendations, the system focuses on controlled content extraction, classification, and recommendations from approved resources.

Human review, traceable outputs, and evaluation tools help administrators assess AI performance, maintain transparency, and identify opportunities for continuous improvement.

Business outcome

Achievion established an intelligent AI foundation that makes the client’s digital health content easier to manage, discover, and personalize while preserving the human oversight required in healthcare environments.

By combining automated content classification, semantic search, recommendation capabilities, and improved analytics infrastructure, the solution creates opportunities to reduce operational effort and deliver more engaging user experiences.

Key business benefits include:

  • Reduced manual content management: AI-assisted protective-factor extraction and tagging streamline repetitive classification tasks while keeping final approval with staff.
  • Greater value from existing content: Semantic search and timestamp-based retrieval make previously difficult-to-find video content more accessible and reusable.
  • Improved content discoverability: Guided navigation and intelligent search help users locate educational resources more efficiently.
  • More personalized user experiences: Context-aware module recommendations create opportunities to increase engagement and improve content completion.
  • Stronger AI governance: Approved content, controlled AI outputs, human oversight, and auditability support responsible AI adoption in a sensitive healthcare environment.
  • Greater ownership of analytics: In-house tracking capabilities provide a foundation for richer user insights, reporting, and future billing workflows.
  • Scalable foundation for digital therapeutics: The reusable AI architecture supports continued platform development and the client’s longer-term vision for clinically effective, potentially reimbursable digital interventions.

The project positions the client to evolve from a content-driven digital health platform toward a more intelligent, personalized, and data-informed digital therapeutic experience.

Timeline 

1 Month
Design Phase

Achievion began by evaluating the client’s existing platform, content management processes, and opportunities for AI integration.

Key activities included:

  • Conducting business and technical discovery sessions.
  • Reviewing content structures, taxonomy, and data workflows.
  • Designing the proposed AI architecture and integration approach.
  • Defining implementation priorities and estimating development effort.
2 Months
Proof of Concept Development

The proof-of-concept phase focused on validating AI-powered content processing and retrieval capabilities.

Key activities included:

  • Developing transcript processing and protective-factor extraction workflows.
  • Testing LLM-based content classification.
  • Implementing initial vector search capabilities.
  • Building a content search interface.
  • Creating an evaluation dashboard to assess AI performance.
4 Months
MVP Development

Following concept validation, the team expanded the solution into an integrated AI-powered platform foundation.

Key activities included:

  • Hardening AI processing pipelines for production use.
  • Integrating core AWS services and data storage.
  • Developing semantic search and guided content discovery.
  • Implementing personalized recommendations and reminders.
  • Building supporting application components and integrations.
  • Testing and refining AI functionality.
2 Months and Ongoing
Platform Integration and Optimization

The ongoing development phase focuses on embedding AI capabilities more deeply into the client’s workflows and expanding analytics functionality.

Key activities include:

  • Integrating AI-assisted content management and administrative search.
  • Expanding personalized recommendation capabilities.
  • Embedding semantic search into the existing platform.
  • Implementing client-side and server-side activity tracking.
  • Developing data schemas and analytics integrations.
  • Performing testing, documentation, and knowledge transfer.

Team

Product Manager 
Solutions Architect 
AI/ML Engineer
Full-Stack Developer 
MLOps Engineer
QA Engineer

Tech Stack

AI

Amazon Bedrock

Application Layer

Python Django
Wagtail CMS

Desktop Content Creator Application
Streamlit Evaluation Dashboard

Data Layer

Amazon RDS
Amazon S3

Amazon S3 Vectors
Amazon OpenSearch Serverless

Infrastructure

AWS Batch
AWS Lambda
Amazon ECR
AWS App Runner
Amazon CloudWatch

Get in touch to move your AI transformation
or product idea from concept to execution.

    *

    *

    0 from 500