DataHex for investment signals
Leverage AI workflows to test thematic investment strategies
Turn unstructured data at scale into the signals that underpin or tilt investment portfolios. AI-native workflows help investors, researchers, and index teams develop, back-test, and publish thematic portfolios, indices, and market-trend insights, moving from theme to validated signal quickly.
Leverage AI workflows to test thematic investment strategies
- 01 · THEMEPlain language to scored universe
- 02 · EVIDENCEVerbatim filing quotes on every score
- 03 · GUARDRAILSPoint-in-time, cost-estimated, governed
- 04 · WEALTHStakeholder capitalism principles
From theme to validated signal
What investment teams get from DataHex Thematic Studio
Benefit from capabilities across the DataHex platform, configured to your use case.
Create, test and verify quickly
New investment themes, thematic portfolios or custom index products.
Full exposure to a theme
Captured with subjective bias removed, so classification stays objective.
Score with confidence
AI guardrails in place to ensure integrity, repeatability and traceability.
Faster time to signal
At lower cost, with run costs known before execution.
Evidence-backed signals
Credible and defensible to investors and regulators.
Wealth creation, independent of price
A screening and weighting input not derived from share price.
Guardrails
Powerful AI, bounded by the right controls
AI can read thousands of filings, transcripts and news items in hours. Without the right data engineering and controls, it can also learn from information that was not available at the time, run up excessive processing costs, and produce scores no one can explain. DataHex builds the guardrails into the workflow.
Point-in-time integrity
Every document is indexed to the date it became public, so a backtest only reads what was knowable at each rebalance. This prevents forward learning and look-ahead bias.
Chunking and context indexing
Source documents are chunked and indexed by company, date and section, so the model retrieves only the passages relevant to each criterion rather than processing whole corpora.
Entity tagging
NLP entity tagging links every passage to the right company and instrument through the security master, so evidence lands on the correct name.
Pre-estimated run costs
Processing costs are estimated before a run executes, so teams choose universe size and depth knowingly.
A bounded workflow
A guided workflow, not an open-ended chatbot. It outputs a thesis to review, refine and approve before anything is scored.
Evidence and audit
Every score is backed by verbatim quotes validated against the source text, with versioned methodology and a full audit trail.
DataHex Thematic Studio
From theme to signal, in one workflow
Thematic Studio scores a market’s companies for their exposure to a theme, grounds every score in verbatim primary-source evidence, and carries it through construction, backtesting, deployment and monitoring. It indexes the corpus you choose: company announcements, earnings transcripts, news and social media.
- 01
Define
Build the thesis in plain language, criterion by criterion, or let the AI assistant draft it from your description and research.
- 02
Score
Score the universe you nominate, from the top 10 to the top 300 or more, on four exposure dimensions.
- 03
Evidence
Every score traced back to the sentence in the filing that justified it.
- 04
Backtest
Weighting and construction rules applied, then performance and risk backtested before committing capital.
- 05
Deploy
Run in production on a schedule and delivered through a point-in-time API, as an index or an alpha signal.
- 06
Monitor
Exposure re-assessed on every new filing, with drift tracked against your thresholds.
Four exposure dimensions
Direct business
Revenue and operations placed squarely in the theme, confirmed against the filings rather than assumed.
Strategic alignment
Stated intent, capital allocation and roadmap: a company moving into the theme before the revenue shows it.
Indirect exposure
Suppliers, infrastructure and enablers one step removed from the theme, which a sector code will never show.
Market perception
How the market already prices the company against the theme, and where the narrative runs ahead of or behind the disclosure.
Difficult decisions, answered quantitatively
The questions Thematic Studio helps answer
Thematic investing turns on judgment calls that are hard to defend. Thematic Studio turns each one into a measured, evidenced answer.
| The investment question | How Thematic Studio answers it |
|---|---|
| Which companies are genuinely exposed to this theme? | A structured, versioned thesis rubric applies the same explicit criteria to every company, scored consistently across the whole universe. |
| How exposed, and in what way? | Four exposure dimensions separate direct revenue from strategic intent, value-chain position and market perception. |
| Can we defend each name to an investment committee or regulator? | Verbatim filing evidence, validated against the source, sits behind every inclusion. |
| Would this theme have worked historically? | Theme to scored universe to constructed, backtested index in one workflow, on point-in-time data. |
| Has a constituent drifted away from the theme? | Exposure is re-scored on each new filing, so drift is a measured quantity with defined thresholds. |
| Does this work outside English-language markets? | A model-agnostic, multilingual engine reads filings in other languages as readily as English. |
Move from theme to validated signal
A walkthrough tailored to your use case, themes and investment universe.