Hub Portfolio Manager for Private Equity

A private equity firm's view of its own portfolio is assembled by hand, every reporting cycle, from sources that disagree. The CRM holds the relationship, a financial database holds the market comparables, portfolio companies send numbers in whatever template they use, and an analyst reconciles all of it into a spreadsheet that becomes the source of truth until the next quarter, when it is rebuilt. Every artefact the firm produces — a tear sheet, a rebalancing analysis, an LP report — is derived from that spreadsheet, which means the firm's external credibility rests on a file with no lineage.

• Data arrives in inconsistent formats from CRMs, financial databases and portfolio companies.
• The reconciled view is rebuilt each cycle, so there is no continuous record and no lineage.
• Reporting artefacts are hand-assembled, which makes them slow and makes errors expensive.
• Cross-functional access is all-or-nothing, so collaboration is limited by what can safely be shared.

Hub consolidates disparate sources into a cohesive ontology built for the private equity sector — CRMs, financial databases such as Preqin, and third-party and open-source data — transforming inconsistent formats into one standardised model. Data quality is enforced at ingestion with built-in validation and governance controls, and generative extraction techniques are used to minimise inconsistencies and validate incoming information. On that base, the investment ontology supports portfolio monitoring, analytics and reporting, accessible both inside the platform and externally through Excel or other systems, with fine-grained permissions that make cross-functional collaboration safe rather than blanket. Applications are then deployed on top: automated proactive reporting, disclosure management, memo creation, and highly personalised output artefacts such as tear sheets, rebalancing analyses and LP reports — with low-code tooling so users can build new metrics, ratios, forecasts or valuation models without waiting on engineering.

The normalisation step is where a model changes the economics. Portfolio company reporting arrives as documents, not feeds, and extracting a consistent set of figures from inconsistent templates is a reading task that has historically consumed analyst time by the week. Generative extraction with validation against the ontology does that work and flags what it cannot reconcile. Everything downstream is deliberately deterministic: reports are generated from the structured model rather than from prose, which is what keeps the output defensible when it is going to a limited partner.

SOURCES

CRMs, financial databases, third-party and open-source data

MODEL

Standardised private-equity investment ontology

QUALITY

Built-in validation and governance at ingestion

ACCESS

In-platform and external (Excel and others), fine-grained permissions

OUTPUTS

Tear sheets, rebalancing analyses, LP reports, disclosure and memos

EXTENSIBILITY

Low-code metrics, ratios, forecasts and valuation models

1. 01Disparate sources normalised into one standardised private-equity ontology.
2. 02Generative extraction used to minimise inconsistency and validate incoming information.
3. 03Portfolio monitoring, analytics and reporting from a single continuous model.
4. 04Fine-grained permissions enabling cross-functional collaboration without blanket access.
5. 05One-click deployable applications for reporting, disclosure management and memo creation.
6. 06Low-code tooling so new metrics and valuation models are built without engineering time.

One model

Behind every artefact the firm producesdesign intent

At ingestion

Validation, against reconciliation each cycledesign intent

• FINANCE AUTONOMYLedger→
• AI-NATIVE BUSINESS PLATFORMNexus→
• PROCURE-TO-PAYHub for Accounts Payable Automation→