At a glance
What was built, on what, and why it matters.
- Industry
- Entertainment & Tourism
- Domain
- Workforce Support & Employee Self-Service
- Solution
- Conversational AI · Employee Knowledge Assistant · Workforce Automation
- Core technologies
- Large Language Models (LLMs) · Python
- Delivery model
- Four-week proof of concept
Key outcomes
Measured where it counts.
- 01Accurate, policy-grounded responses through a conversational interface
- 02Met proof-of-concept accuracy benchmarks
- 03Reduced reliance on manual support for routine policy enquiries
- 04Established a foundation for broader enterprise AI adoption
6,000 enquiries, every month.
The organisation needed a better way to give employees workforce policy information — many of whom had no access to internal systems.
Support teams handled around 6,000 policy enquiries each month, creating overhead and bottlenecks. An existing chatbot couldn’t consistently give accurate, contextual responses, limiting confidence. They needed reliable, policy-grounded answers with room to expand into other employee services.
Reliable answers, policy-grounded.
LLMs + prompt engineering, validated against success criteria.
DBiz designed and implemented a web-based AI assistant focused on workforce leave-policy enquiries.
Using large language models and advanced prompt engineering, employees ask natural-language questions and receive concise responses grounded in approved policy documentation. Delivered as a four-week proof of concept, it included authentication, persona design, QA processes, stakeholder validation, and automated testing — with success criteria for accuracy, contextual relevance, and citation quality.
Confidence, and a path forward.
The assistant delivered accurate, policy-grounded responses through a self-service experience tailored to workforce support. Iterative refinement and validation resolved the accuracy and context issues of the previous implementation.
The successful proof of concept increased confidence in AI adoption and established a path to extend self-service to additional business functions.
Technology stack
| Layer | Technology |
|---|---|
| AI platform | Large Language Models (LLMs) |
| Knowledge sources | Approved Policy Repositories · Designated External Sources |
| Quality & optimisation | Prompt Engineering · Stakeholder Validation |
| Test automation | Python |
| Delivery model | Web-Based AI Assistant |