AI Scoring Methodology

How Scoring Works

When you click “Score Deal” on any deal in your portfolio, we send the deal’s structured data to an AI model (Claude by Anthropic). The model evaluates the deal across four dimensions and returns a score for each, along with the specific factors that drove each rating up or down.

The overall score is a weighted average of the four dimension scores:

Scoring Dimensions

Score Scale

Both the overall score and each dimension score use the same 0–100 scale:

Confidence Indicator

Every score displays a confidence level based on how complete the underlying data is. More populated fields mean higher confidence.

Data Sources

The scoring model uses only the structured data you have entered into SyndTrack for each deal:

• Deal details — name, asset class, location, committed and called capital, target IRR, preferred return, waterfall position, hold period, and vintage year.
• Sponsor profile — display name, total AUM, years in business, number of deals, average IRR, and average equity multiple.
• Transaction history — total capital calls, total distributions, and transaction count.
• Projections — projected IRR and projected MOIC, when available.
• Document count — number of uploaded deal documents (documents are not read by the scoring model).

The model does not have access to the internet, third-party databases, or any data outside of what you provide. It does not read the contents of uploaded documents.

Known Limitations

• Scores reflect data quality — If deal data is incomplete, the model assigns neutral scores (40–50) for missing dimensions. A low score may reflect missing data rather than a poor deal.
• No real-time market data — The model does not access live interest rates, cap rates, or comparable sales. Market condition scores are based on asset class and location only.
• AI model variability — Scores may vary slightly between runs as AI models are updated. Historical scores are preserved so you can compare over time.
• Single-deal scope — Each deal is scored independently. The model does not consider portfolio-level diversification or correlation between your deals.
• No qualitative analysis — The model cannot evaluate sponsor reputation beyond the numbers you provide. It does not account for personal relationships, legal disputes, or news events.

What the Score Does NOT Capture

The AI deal score is a structured analysis of the data you provide. It is not a recommendation and should not be the sole basis for any investment decision. Specifically, the score does not account for:

• Your personal financial situation, risk tolerance, or investment goals
• Tax implications specific to your jurisdiction or entity structure
• Legal terms in offering documents beyond what is captured in structured fields
• Sponsor character, communication quality, or investor relations track record
• Macroeconomic forecasts, geopolitical risk, or regulatory changes
• Correlation with your existing portfolio or concentration risk

When the Score Is Useful

• Comparing deals side by side — Scores provide a consistent framework for evaluating multiple opportunities on the same criteria.
• Identifying data gaps — Low confidence or low dimension scores often highlight missing information you should request from the sponsor.
• Tracking changes over time — Re-scoring a deal after new data (distributions, valuations) shows how the deal is performing against expectations.
• Starting due diligence — The factor breakdown highlights areas that warrant deeper investigation before committing capital.

Important Disclaimer

Questions

If you have questions about how scoring works or want to report an issue with a score, contact us at support@syndtrack.io.