Exploring GenAI for Social Impact at the AWS Breaking Barriers Hackathon

This month, our engineers took part in the AWS Breaking Barriers Hackathon at Manchester’s Bridgewater Hall. The event brought together engineers, AWS specialists and charity partners, including The King’s Trust, Cancer Research UK and Race Against Dementia, to tackle some of the biggest challenges charities face today - from limited resources and funding instability to outreach and impact measurement.

Over three days, teams collaborated closely, exploring how generative AI can drive practical, real-world change across healthcare, sustainability and social impact.

Tackling Youth Unemployment with GenAI

Our engineers worked with The King’s Trust, focusing on building solutions to support their mission to tackle youth unemployment and make their programmes and resources more accessible to young people. Two cross-functional teams delivered fully working GenAI-powered platforms:

AI-Guided Career Pathways with Personalised Roadmaps

This winning solution in Manchester helps individuals discover careers that align with their skills and interests, generate personalised roadmaps, and access relevant King’s Trust courses and career events.

The team built an AI-powered career-matching quiz that:

• Matches users to careers aligned with their skills and interests
• Generates a personalised career roadmap based on responses
• Returns relevant King’s Trust courses and careers events tailored to the individual

The solution used AI agents built in Amazon Bedrock, which were prompted to pull data from S3-based knowledge bases containing information on King’s Trust programmes and UK careers events.

At its core, the recommendation engine used a Bayesian-based methodology - an adaptive decision model that refines itself as users progress through the quiz. This allowed recommendations to become more accurate based on user input, rather than relying on static rules.

The solution directly supports The King’s Trust’s mission to recognise and tackle youth unemployment.

A Serverless, AI-Powered Education Hub for Delivery Partners

The second team focused on reducing the friction delivery partners face when trying to find, evaluate and use King’s Trust programme resources.

Our engineers built a serverless, AI-powered Education Hub, with Amazon Bedrock at the core, designed to make content easier to discover and simpler to navigate.

‍

The platform introduced four key improvements:

• Semantic search (natural language): Delivery partners can search in plain English — they don’t need to know internal King’s Trust terminology or where a resource sits in the site structure. The search updates results dynamically and returns the most relevant resources.
• AI summaries (reduce guesswork): Before downloading anything, users can preview a short, accurate AI-generated summary of what’s inside a resource, so they can confidently choose the right file first time.
• Clearer navigation (unified hub): They provided a single, consistent place to browse content, organised logically by programme → unit → activity, with filtering that’s driven by metadata and supported by AI to improve findability.
• Trust Coach (in-app assistant): An embedded assistant answers content questions instantly and suggests the most relevant resources, helping users move faster without having to hunt through the library.

Together, these improvements reduce friction and directly support the impact goals: more usage, more reach, and more downloads - from both existing and new delivery partners.

‍

A big thank you to AWS, The King’s Trust, Cancer Research UK, Race Against Dementia, and everyone involved for creating the space to build, collaborate, and focus on real-world impact.