Steps to Prepare Your Data for AI in UniPhi

Maximise UniPhi’s Own Modules

Use UniPhi’s built-in modules to capture as much operational and project information as possible. The more complete your records, the better the AI output.

Import Data from Other Systems

UniPhi offers multiple integrations. If you use other tools such as Microsoft Project, Primavera P6, or even Excel spreadsheets of past projects, you can import that information directly into your UniPhi deployment. This can be historical project data, cashflow profiles, or resource allocations, all of which can train AI algorithms for better forecasting.

Capture Comprehensive, Multi-Faceted Data

Aim to record not only time, cost, and quality metrics, but also:

Conversations and Sentiment: Import emails and comments into UniPhi (e.g., via the Outlook add-in) so that AI can assess tone and detect early warning signs in stakeholder communications.

Metadata: Capture details such as client, supplier, sector, project type, and location, these enrich AI models and improve their relevance to your specific context.

Operational Detail: Include issue logs, RFIs, design changes, and defect records. These form the raw material for future risk predictions and lessons learned analysis.

Commit to Data Quality

Missing or misleading data can undermine AI-driven insights. For example, if a project has “no risks” flagged when in reality risks exist, AI will detect anomalies and may flag data quality issues. UniPhi’s algorithms are difficult to “game” because they depend on real, up-to-date inputs such as progress claims and time status updates.

Why This Matters Now

UniPhi 21 introduces AI-powered features like the AI Summariser, which uses OpenAI’s ChatGPT to instantly produce executive summaries of issues, documents, or meeting notes directly within the platform. These tools save time, in internal testing, they’ve cut the review process for complex issues from minutes to seconds, but they only work well if the underlying data is comprehensive and up to date.

The future of AI in UniPhi goes far beyond summarisation. Planned developments include:

• AI-assisted project estimating using historical performance data

• Predictive risk profiling

• Vendor shortlisting based on performance history

• Automated proposal drafting from prior project data

• Resource forecasting for senior/junior roles based on past workloads

  
  

All of these rely on having your data in order now, whether through consistent use of UniPhi’s modules, importing from other systems, or setting up integrations to centralise information.