Structuring the Unstructured
Fifteen years ago, one of Australia’s best investors told the Calibre team: “Alpha is simply the ability to compare.”
To outperform the market, fundamental investors need a repeatable way to assess mispricing across diverse opportunities. But the reality of modern investment research is that the vast majority of valuable information – earnings call transcripts, 300-page annual reports, auditor footnotes, expert network interviews, even your proprietary internal research – is entirely unstructured.
For decades, the holy grail of investment technology has been converting this qualitative chaos into structured, comparable data. Excel did it for financial numbers. Research Management Systems did it for note-taking and compliance.
Now, artificial intelligence is driving the next great leap forward. Today, Calibre announces the release of Intelligent Scorecards within CalibreRMS Intelligence.
Beyond Summarisation: From Text to Structured Data
Intelligent Scorecards represent an evolutionary step beyond AI summarisation. Previous features like Intelligent Notes allowed analysts to instantly transcribe conference calls or summarise earning announcements or Annual Reports into formatted text, but the output remained fundamentally text. You cannot chart a text summary, run a quantitative screen on a paragraph, or aggregate qualitative takeaways across a 50-stock portfolio.
Intelligent Scorecards solve this problem. By pairing frontier AI models (GPT-5.5, Gemini 3.5, Claude 4.8 Opus) with Calibre’s configurable scorecard architecture and proprietary time-series database, analysts can automatically extract, categorise, and score unstructured data into rigid structures: numerical values, 1–5 rankings, or categorical drop-downs (Weak / Average / Strong).
Because these scorecards are stored in a time-stamped database, qualitative insights instantly become quantitative data points. Tracked, charted, screened, and aggregated over time.
The intelligence behind each scorecard lies in its custom prompts. Every Intelligent Scorecard template contains team-defined and crafted prompts, also known as skills, that instruct the AI exactly how to interpret source content and score each field. These skills encode your firm’s unique analytical framework: defining what constitutes a “red flag,” how to weight conflicting signals, and where to set thresholds for categorical ratings. A prompt might instruct the model to flag insider selling only when it exceeds $1 million within 30 days of negative guidance, or to classify capital allocation as “Poor” when acquisitions exceed 50% of free cash flow for three consecutive years. Because these skills live inside the team scorecard template, every analyst on your team applies identical criteria to every company, eliminating subjectivity and ensuring institutional consistency at scale.
The Analyst Tool: 6 Ways Intelligent Scorecards Extract Alpha at the Company Level
When analysing a single stock, the analyst’s job is to read between the lines of corporate disclosures. Intelligent Scorecards act as a tireless junior associate, processing thousands of pages and grading them against your team’s proprietary criteria using the team skills.
…
1. Systematic Detection for Forensic Accounting
Corporate failures rarely happen without warning, they’re usually preceded by forensic accounting anomalies. Configure Intelligent Scorecards as a systematic detector that scans every annual report and disclosure:
Earnings Manipulation: Extract operating cash flow and net income, flagging risk when net income is highly positive while operating cash flow is deeply negative.
Auditor Anomalies: Calculate audit fees as a percentage of revenue, flagging large-cap companies using unknown micro-auditors or paying abnormally low fees.
Insider Selling Disconnects: Correlate insider trading disclosures with recent price action, assigning a Risk Rank (1–5) when executives dump stock after material price collapses.
“Zombie” Debt Traps: Extract interest coverage ratios, flagging companies generating just enough cash to cover interest but unable to pay down principal.
.
.
2. Decoding Executive Compensation Alignment
Remuneration reports are notoriously dense, often burying bonus hurdles in complex legalese. Drop an annual report into CalibreRMS, and the AI evaluates compensation structures:
Numerical: Extract the exact ROIC hurdle rate required for the CEO’s long-term incentive plan.
Categorical: Evaluate alignment with shareholder interests, selecting from pre-defined options: Strongly Aligned, Neutral, or Poorly Aligned.
.
.
3. Unearthing Footnote Red Flags
The most critical risks often hide in the “Notes to the Accounts”: related party transactions, off-balance-sheet liabilities, revenue recognition changes. The AI scans filings and assigns a Risk Rank (1–5) based on severity, while populating a text field summarising the specific concern (e.g., “Auditor flagged material uncertainty regarding a related-party loan extension”).
.
.
4. Grading Management Credibility
Earnings calls are rich with tonal shifts and subtle language changes. Feed call audio into CalibreRMS to populate a “Management Credibility Scorecard.” By comparing the current transcript against four quarters of prior guidance, the AI scores consistency and sentiment. When management previously promised margin expansion but now blames “macro headwinds”, the AI assigns a lower credibility score.
.
5. Standardising Environmental and Social Metrics
Obscure environmental statistics, Scope 3 emissions, water usage, community engagement programs etc, are often buried in unstandardised sustainability reports. The AI locates exact figures, normalises them, and inputs them as numerical values. Using Calibre’s integration with live-web search tools, scorecards can also evaluate real-world social risks (supply chain controversies, labour disputes) and categorise Social Risk as High, Medium, or Low, complete with cited news links.
.
.
6. Dynamic Excel Model Integration
The most powerful feature of Intelligent Scorecards is what happens after the data is structured. Because outputs live in Calibre’s time-series database, analysts can link quantified AI outputs directly into financial models via the Calibre Excel Add-in.
When a Red Flag Scorecard detects severe footnote risks, or an Executive Compensation Scorecard ranks management as “Unaligned,” this data feeds dynamically into Excel DCF models. Program your model to automatically increase WACC by 150 basis points or lower terminal growth rates to reflect heightened governance risk. Qualitative observations converted to standard signals can now systematically influence quantitative valuations.
.
The Portfolio View: 5 Ways Intelligent Scorecards Transform Idea Generation and Portfolio Management
The true power of structuring the unstructured emerges at the universe and portfolio level. Because Intelligent Scorecards save data into Calibre’s time-series database, these custom metrics can screen and evaluate your entire coverage universe.
.
1. Universe Screening Beyond Financial Metrics
Traditional fundamental screens are limited to financial data. With Intelligent Scorecards, screen on proprietary qualitative insights you can now run a filter across 1,000 stocks to exclude any company with six or more “Red Flags” across forensic accounting and governance scorecards. You’re no longer screening on market consensus, you’re side-stepping landmines or finding aligned management teams using proprietary insight, executed at scale.
.
.
2. Portfolio Aggregation vs. Benchmark
Portfolio Managers need visibility into aggregate risk. If analysts capture “Supply Chain Risk” (1–5) via Intelligent Scorecards, PMs can instantly calculate the weighted average score across the portfolio and compare against the benchmark. When your portfolio’s Governance Risk Score averages 3.8 versus the benchmark’s 2.5, you have immediate, structured visibility into an active risk position.
.
3. Time-Series Tracking for Thesis Drift
One of the greatest destroyers of alpha is thesis drift. Holding a stock long after the original investment rationale has deteriorated. Because Intelligent Scorecards are time-stamped, you can track and chart qualitative changes. When the AI systematically downgrades a company’s capital allocation from ‘5’ to ‘2’ over four quarters, the system triggers an automated alert. PMs see the qualitative thesis breaking before the broader market prices in poor management decisions.
.
4. Custom Quantamental Ranking
Quants have traditionally struggled to incorporate fundamental analysts’ qualitative views because the data was trapped in research notes and text documents. Intelligent Scorecards bridge this divide. Extract structured AI outputs via Calibre’s API to build custom ranking models, for example, weighting momentum and value while adding your firm-specific “Footnote Red Flag Score” as a negative overlay.
.
5. Automated ESG and Regulatory Reporting
Regulations mandate rigorous reporting on climate risks and active ownership. When your team uses Intelligent Scorecards to automatically extract Scope 2 emissions (numerical), Board Diversity metrics (categorical), and Engagement Outcomes (ranked) from every company meeting, reporting becomes frictionless. CalibreRMS aggregates these data points to generate compliant, audit-ready reports instantly.
Relying on third party providers for this ESG scoring carries an embedded research process from their own methodology which may not be visible. Now you can replicate this with full process and methodology transparency.
.
The Calibre Advantage: Your Process, Your Edge
The shift from text-based AI summaries to structured AI scorecards represents another shift in investment technology. But the underlying philosophy remains unchanged: public information is becoming an alpha desert.
Buy a pre-packaged ESG score or generic risk rating from a third-party vendor, and you’re buying consensus. Every other fund has access to the exact same data.
Intelligent Scorecards are different because they’re built on your proprietary investment process. You define the forensic red flags. You define what makes executive compensation “Strongly Aligned.” You build the Skills. You decide how scores systematically alter discount rates in your models. The AI simply acts as an infinitely scalable analyst, applying your unique worldview to the mountain of unstructured data the market produces daily.
True to Calibre’s enterprise-grade security architecture, this is achieved via a Bring-Your-Own-LLM model. Your proprietary scorecard structures, prompts, and source documents are processed securely through your own API keys. Your data never trains public models and your intellectual property remains strictly yours.
.
Stop Reading. Start Scoring.
The bottleneck in modern asset management is no longer information availability, it’s human attention. The teams that outperform over the next decade won’t be those who read filings fastest. They’ll be the teams who convert unstructured chaos into rigid, comparable data that drives repeatable, thesis driven investment decisions.
With Intelligent Scorecards and dynamic Excel integration, that capability is now natively embedded in your research workflow.
Turn your team’s qualitative insights into quantitative alpha. Contact Calibre Financial Technology today to schedule an interactive demo of Intelligent Scorecards.
Share on Facebook
Share on X
Share on LinkedIn
