Features & Roadmap

Public officers can easily create their own AIBots by providing essential details such as Bot Name, Agency Name, Bot Description, and Bot Instructions. For high-risk use cases, creators must seek approval before proceeding.

To ensure accurate responses, upload a relevant knowledge base. Since LLMs only understand internet-based data, your bot will benefit from access to your agency’s specific information. Note that tables and images are not well-processed by the bot, so textual documents should be added for clarity. You can also upload content classified up to RESTRICTED \ SENSITIVE NORMAL for reference.

Refine your bot’s performance by configuring system prompts to match your desired responses. Adjust the bot’s personality to influence its tone and style. Experiment with creativity levels—the higher the creativity, the more varied the responses will be. Keep tweaking settings to achieve your ideal chatbot behaviour.

Test your bot by asking common questions and evaluating the answers. Gather feedback from both yourself and real users, then iterate. If the responses need improvement, adjust instructions and test again. After finalising the bot’s performance, submit it for approval and share or collaborate with others across your agency.

• Enable bot analytics and conversation data downloads.
• Integrate additional non-OpenAI models.
• Support multi-modal input capabilities.

• Develop API endpoints to extend bot usage beyond the AIBots platform.
• Launch Agency Admin Portal.
• Enable support for higher classification levels.
• Expand device access and extend beyond WOG network (e.g. MOE students).
• Implement chunk size modification capabilities.

• Enhance agent and actions functionality.
• Implement integration with SharePoint and other applications.