Meet Hudini, Hudu's AI assistant. Hudini is a smart AI assistant that enhances your platform with AI-driven assistance. Configure it with your preferred AI provider to answer questions and generate content, giving users instant access to the information they need. Hudini now also offers KB article awareness, meaning it can recognize the knowledge base article you are viewing and provide a summary or answer questions about it.
Supported providers:
- Open AI
- Azure OpenAI
- Grok
What you will need:
- Admin or Super Admin to toggle Hudini On
- For Open AI, you will need:
- API Key
- Model
- For Azure OpenAI, you will need:
- API Key
- Endpoint
- Deployment Name
- API Version
- For Grok, you will need:
- API Key
- API Endpoint
- https://api.x.ai/v1
- Model
Please note that Grok with Hudini does not currently support function calling like OpenAI.
ℹ️ Note: Hudu does not manage or include OpenAI usage. You will need to connect your own OpenAI account and have API credits or billing set up separately.
Guides
Getting Started
- To start using Hudini:
- An Admin will need to turn it on in Admin > Hudini.
- Once it is turned on, the Hudu AI Assistant will appear on the toolbar in the top right.
- Choose preferred provider.
- Enter required credentials for chosen provider.
- Select Save.
Configuring Azure OpenAI
What Hudu needs from Azure OpenAI:
- Endpoint
- API Key
- Deployment name
In Azure:
- Navigate to Azure Open AI
- Click +Create or + Create Azure OpenAI
- Click Next
- Configure network security for your Azure AI services resource
- Choose a type of network security that works for your organization
- Click Next
- Add Tags (optional)
- Click Next
- Review and Submit
- Make sure everything looks good
- Click Create
- Click Go to Resource
- Select Explore Azure AI Foundry Portal
- Select Create new deployment
- Choose From base models
- Select a Model
- Click Deploy
Navigate to Shared Resources > Deployments on the left side panel to copy your API Key and find your endpoint.
In Hudu:
- Navigate to Admin -> Hudini
- Enable Hudini
- Find and select Azure OpenAI from the list
- Input your API Key and Endpoint from your Deployment
- Please note that the endpoint provided will need to be edited to look like the following (deployment name should be the same as the model selected for the deployment):
- Enter API Version:
- The version can be found at the end of the Target URI in the Endpoint box
- This example may be different than the version at the end of your URI: example: 2023-05-15
- The version can be found at the end of the Target URI in the Endpoint box
Setting up Prompt Rules
Define pre-made prompt rules that users can select when interacting with Hudini. A well-crafted set of prompt rules helps improve the quality and consistency of AI-generated responses.
- To add a new prompt rule:
- Select New Prompt Rule.
- Name the Prompt.
- Input the body of the prompt.
- Click Create to save the prompt
- Prompts will appear in a drop-down in the Hudu AI Assistant to select from.
🔗 This assistant is designed with safety and ethics in mind. Learn more in our Hudu AI Principles article.
Hudini Prompt Examples
Context: You are assisting in structuring Quick Notes for a company’s dashboard. These notes are entered into a rich text WYSIWYG editor and should follow a standardized, easy-to-read format.
Instructions:
- Format the note in a clear, structured layout for quick scanning.
- Use bold headers to clearly separate sections.
- Use bulleted lists or short, well-formatted lines for readability.
- Keep details concise and relevant to the client’s current state.
- If any information is missing, use [N/A] instead of leaving blank.
- If any content starts with "A side note:", format it as a Warning Callout.
Context: You are provided with raw, unstructured technician notes. Convert them into a clear, concise, and easy-to-follow Standard Operating Procedure. This SOP should be suitable for a junior technician.
Instructions:
- Begin with a brief, descriptive title that reflects the task.
- Structure the SOP using numbered steps.
- Clarify vague language and remove redundancies.
- Format for readability using spacing and consistent punctuation.
- Ensure terminology is accurate and internally consistent.
- If a step includes warnings or caveats, clearly highlight them.
- Input: [insert technician notes here]
Context: Review the provided documentation and identify any missing or unclear elements that would prevent another technician from successfully completing or understanding the process.
Instructions:
- Identify incomplete sections or vague instructions.
- Suggest any missing prerequisites, steps, tools, or context.
- Ensure terminology, access details, and outcomes are clearly stated.
- Output should be structured as a list of documentation improvement recommendations.
- Input: [insert documentation content here]
Context: You are provided with rough meeting notes. Convert these into a structured summary of action items.
Instructions:
- List each action item clearly with the assigned person.
- Include due dates if mentioned.
- Identify and call out any blockers.
- Use clear, bullet point formatting for quick scanning.
- Ensure the tone remains professional and objective.
- Input: [insert meeting notes here]
Context: You are summarizing a recent support call or ticket for handoff to another technician or team.
Instructions:
- Issue: Describe the core problem.
- Actions Taken: Detail what has been attempted or resolved.
- Client Response: Note the client's reaction or instructions.
- Next Steps: Outline what still needs to be done, or who it's handed off to.
- Keep the format clean and consistent.
- Be factual and avoid assumptions unless noted as such.
- Input: [insert raw call or ticket notes here]
Context: You are generating a structured knowledge base article for internal or client-facing use.
Instructions:
- Begin with a short but descriptive introduction.
- Break down the procedure into clear steps with numbered or bulleted formatting.
- Use section headers where appropriate (e.g., Overview, Steps, Troubleshooting, Additional Notes).
- Include common issues or troubleshooting advice if relevant.
- Use language appropriate to the audience (technical or non-technical).
- Input: [insert topic or draft content here]
Context: Before proceeding with the request, identify and remove sensitive customer information. Replace these details with appropriate placeholders.
Instructions:
- Search for and replace the following:
- Company names → [Company Name]
- Employee names and emails → [Full Name], [Email]
- Phone numbers → [Phone Number]
- Physical addresses → [Address]
- IP addresses → [IP Address]
- API keys or credentials → [Credential/Key]
- Unique customer identifiers → [Customer ID]
- Maintain overall clarity and integrity of the request.
- Output the sanitized version only.
- Input: [insert raw content here]
Context: You are creating a troubleshooting guide for an MSP technician.
Instructions:
- Start with a brief summary of the issue.
- List possible causes in a bulleted or numbered format.
- Provide step-by-step solutions for each cause.
- Include preventative tips or ongoing monitoring suggestions at the end.
- Ensure formatting is clean and language is concise and instructional.
- Input: [insert description of issue or scenario here]