Internal AI Search Assistant

Overview

A proof-of-concept conversational search experience designed to help colleagues discover internal knowledge and workplace information more easily, exploring natural-language search instead of manual intranet navigation.

Context

Important employee information can be distributed across different intranet pages, systems and internal resources, making it difficult for colleagues to quickly find the information they need.

Challenge

Explore a simpler way for colleagues to find relevant internal information without needing to know exactly where it is stored.

Idea

Use a conversational AI search assistant where colleagues can ask questions using natural language, rather than navigating through department pages, documents and policies to find an answer.

Experience

The concept explored providing direct access to information across areas such as HR, payroll, working hours, annual leave, payslips, invoices and internal opportunities, without claiming every source was fully integrated.

Accessibility

Designed with keyboard accessibility, semantic HTML, clear focus states, high colour contrast and screen-reader-friendly labelling in mind, so the assistant could supplement normal navigation rather than replace it inaccessibly.

My Contribution

• Built a working proof-of-concept AI search experience focused on improving access to internal knowledge resources.
• Developed the chatbot UX and search-interface design for the concept.
• Connected the OpenAI Responses API to a vector-store-backed sample knowledge base.
• Built the proof of concept with accessibility as a core consideration.

What I Learned

A proof-of-concept exploring how AI-powered conversational search could improve the discoverability and accessibility of internal knowledge resources.