Leveraging the UniPhi 21 AI Summariser

Introduction to the AI Summariser

UniPhi 21 introduces the AI Summariser, powered by OpenAI’s ChatGPT, directly within the platform. This feature provides quick, context-aware summaries of documents, issues, and comment threads, without needing to copy and paste text into external tools.

By condensing lengthy reports or complex discussions into concise, actionable overviews, the AI Summariser can save significant time and improve decision-making.

Key Concepts

Contextual Summarisation: The AI summarises content based on its context within UniPhi. For example, it can instantly create an executive summary from a PCG report.

In-Memory Processing: Comments and content are passed to OpenAI’s ChatGPT, processed in-memory, and never stored externally.

Immediate Availability: The AI Summariser is available from day one in UniPhi 21.

Steps to Use the AI Summariser

1. Navigate to the Relevant Content

   
   

   • Documents module: Summarise reports, proposals, or other documentation.

   • Issues module: Summarise long-running issue threads or discussion histories.

   
   

2. Locate the AI Summariser Button

   
   

   Look for a new icon (similar to stars or moons, a common design for AI integrations) within the relevant record.

   

3. Click to Summarise

   
   

   Select the button, and UniPhi will process the text and return a concise summary.

   
   

   • Example: In an issue with multiple comments, the AI might highlight the main concerns, assigned actions, and any external references mentioned.

     
     

4. Review and Interpret the Summary

   
   

   • All summaries display the note “AI summary may contain mistakes”.

   • A Read More link explains how the feature works, how to interpret its output, and the source of the AI service.

   • The AI can extract contextual details, such as who is assigned to an action, from the surrounding project data.

   

5. Help Train the AI with Feedback

   
   

   • If the summary is accurate and useful, click Thumbs Up.

   • If it’s unclear or misleading, click Thumbs Down.

   • This feedback loop improves future performance.

Real-World Example

A project manager is brought into a long-running infrastructure project issue they haven’t been involved in before. The issue record in UniPhi contains over 80 comments from multiple stakeholders, with a mix of technical details, meeting notes, and follow-up questions.

Without the AI Summariser, the project manager would have to read through every comment to understand the background, a process taking 10–15 minutes.

By clicking the summariser button, they instantly receive a concise overview highlighting:

• Two main concerns: a delay with a key supplier and a design defect discovered during excavation.

• Actions already assigned, including who is responsible for each.

• Related documents and references for further review.

  
  

In under a minute, the project manager is fully briefed and can act without losing momentum.

Important Considerations

Token Purchase Required

Using the AI Summariser requires purchasing usage tokens. Pricing will be confirmed closer to the UniPhi 21 launch.

Data Dependency

The summariser is most effective if you actively log issues, append conversations via the Outlook add-in, and write comments directly in UniPhi. Without rich in-system data, the summariser has less to work with.

Future Expansion

While the AI Summariser is currently available in the Documents and Issues modules, UniPhi plans to extend it to other areas in the future, potentially including timesheets and communications.