Ask Bobby

The Problem

Finding LGBTQ+ support in the UK can be surprisingly difficult. Information is scattered across charity websites, local authority pages, social media groups, and out-of-date directories. People often need help at the exact moment they have the least energy to search properly.

The language barrier matters too. Someone may not know whether they need counselling, community connection, chemsex support, trans-inclusive healthcare, domestic abuse advice, or simply somewhere safe to talk. A normal keyword search expects the user to already understand the support landscape.

Ask Bobby was created to make that first step easier: describe what is going on in plain language, and be guided towards real organisations, services, and resources that have been checked by people.

The Product

Ask Bobby is a conversational directory, not a standard chatbot. It helps users translate a real-life situation into useful signposting while keeping the experience calm, accessible, and community-aware.

• Plain-language search — users can ask for help in their own words, including slang, uncertainty, or fragmented descriptions.
• Human-verified records — listings are reviewed before appearing, so the system is grounded in real services rather than open-web guesswork.
• Community-specific context — the product understands LGBTQ+ support needs that mainstream directories often flatten or miss.
• Accessible first step — the interface is designed for people who may be anxious, overwhelmed, newly out, isolated, or in a moment of crisis.

Safeguarded AI

AI is useful here because it can understand messy human language and reduce friction. It is also risky if it starts inventing advice, organisations, or certainty. Ask Bobby is therefore built around controlled AI: the system works from checked information, uses guardrails, and routes users towards real-world support rather than pretending to be the support itself.

This is the core of my HCMA approach in practice: AI assists with interpretation and accessibility, while humans define the service boundaries, verify the knowledge base, and keep safeguarding at the centre of the product.

Human Knowledge, Not AI Guesswork

The strongest part of Ask Bobby is not the model. It is the structure around it. Organisations are captured with their real context, eligibility, geography, boundaries, and contact routes. Information is written and checked by humans, then surfaced through a conversational layer that makes it easier to reach.

That distinction matters. A vulnerable user should not be sent to a fictional helpline, an unsuitable service, or a resource that only sounds right. Ask Bobby is built to keep the answer connected to real records and clear next steps.

Privacy and Trust

Ask Bobby handles sensitive intent. Someone may be searching for help around identity, safety, sex, drugs, isolation, abuse, or mental health. The platform is designed with that context in mind: restrained analytics, careful data handling, and a strong line against turning support-seeking behaviour into advertising data.

The product exists to get people to appropriate support, not to profile them.

The Impact

Ask Bobby is signposted to over one million active monthly users, making it one of the most visible LGBTQ+ digital support routes in the UK. It demonstrates what GLYNDA CIC does best: take a messy, high-stakes community need and turn it into a usable, scalable product with safeguarding built in.

The work contributed to Charlie Garnish receiving the Best LGBTQ+ Wellbeing & AI Specialist 2026 award from The IT Awards, recognising the role of community-led AI in reaching underserved groups responsibly.

Why It Shows My Work

Ask Bobby brings together the areas I care about most: community-specific product design, AI that behaves within boundaries, privacy-conscious architecture, live operational data, and a front door that feels approachable to people who may not know where to start.

It is not a demo. It is a live support product serving a real community at scale.