Help, FAQ & Product Guide

What is Helix and who is it for?

Helix is a fund data platform built to give asset managers, fund administrators and operations teams a single, clean, validated view of their fund data across every output they need to produce.

• The mission is to become the single pane of glass for firm data: one trusted data layer that feeds operations, regulatory submissions, distributor templates, marketing and client communications.
• DataHub is the first product in that direction, focused on ingestion, normalisation and distribution of share class and fund data.
• Designed for fund operations teams, compliance officers, data managers and distribution teams at asset managers and fund administrators of any size.
• No minimum fund count and no large implementation project required. You can upload your first file and see mapped outputs within minutes of registering.
• Built to sit alongside existing tools and workflows rather than requiring a full cutover from day one.

What is DataHub?

DataHub is Helix's ingestion and normalisation layer: it takes raw fund data files in any common format, maps the fields to a validated internal schema, and produces clean, standards-compliant outputs ready for distribution or downstream systems.

• Ingest: Accepts CSV, XLSX and PDF (factsheet) files. Multiple files can be uploaded in a single session.
• Normalise: Maps source column headers and values to a consistent internal schema using a rules-first engine with AI assistance for ambiguous or non-standard headers.
• Validate: Applies required-field checks, type checks, cross-field consistency rules, and template-specific rules before export is permitted.
• Export: Produces Openfunds, EMT and EPT outputs in CSV or XLSX format, ready to send to platforms, distributors or internal teams.
• Reduces manual spreadsheet stitching, copy-paste errors and distributor rejections caused by inconsistent field formatting.
• Provides a field-level audit trail so every transformation decision is traceable and reviewable.

What inputs can Helix ingest?

Helix accepts structured and semi-structured fund data files in three formats today.

• CSV: Flat comma-separated files. Column headers are matched to schema fields automatically. Works well for administrator exports and internal data feeds.
• XLSX (Excel): Single or multi-sheet workbooks. Helix extracts the primary data sheet and normalises headers and values in the same way as CSV.
• PDF (text-based factsheets): Helix extracts tables and key identifiers from text-based PDFs. Best results come from clearly labelled fields near visible identifiers such as ISIN and currency.
• Strongly recommended identifiers: ISIN, full share class name, and share class currency. These are required or strongly recommended for most output templates.
• Scanned or image-only PDFs are accepted but produce lower extraction quality. Where possible, use a text-based PDF or re-upload the data as CSV or XLSX.
• Multiple files can be uploaded in a single session, allowing batch ingestion of a full fund range at once.

What outputs can Helix generate?

Helix produces three types of standards-compliant output today, all versioned and aligned to published industry specifications.

• Openfunds v2.1: Share class master data in the Openfunds standard. Covers share class attributes, fees, lifecycle dates, and distribution data. Used for exchange between fund managers, administrators and platforms.
• FinDatEx EMT 4.2 (European MiFID Template): Covers investor profile, target market and cost data required for MiFID II distribution. Required by most European and UK distributors.
• FinDatEx EPT 2.1 (European PRIIPs Template): Covers performance scenarios, PRIIPs cost calculations and transaction cost data. Required for PRIIPs-in-scope products distributed in the EU and UK.
• All exports are available in CSV or XLSX format.
• Outputs are validated before export. A file that fails required-field or consistency checks cannot be exported until the issue is resolved.
• PRIIPs production workflows are supported. Contact us for current scope and availability.

How does mapping work?

Helix uses a rules-first mapping approach, applying deterministic matching before falling back to AI-assisted analysis for ambiguous cases.

• Exact match: Column headers that exactly match schema field names or known aliases are mapped automatically with 100% confidence.
• Pattern match: Common variations, abbreviations and regional naming conventions (e.g. "ccy", "currency", "share_ccy") are matched via a pre-built rule set at 90% confidence.
• AI-assisted match: Where rule-based matching does not resolve a header, a language model analyses the column name and sample values to suggest the most likely schema field. AI-assisted matches are capped at a defined confidence ceiling and flagged for review.
• Exception handling: Unresolved headers are surfaced in the audit trail with a prompt to manually confirm or reassign the mapping before export.
• ISIN and other unique identifiers are given priority weighting in the matching logic to reduce the risk of identifier mis-mapping.
• Manual mapping overrides can be applied in the interface. All overrides are logged in the audit trail.

How do you prevent hallucinations and errors?

Helix is built on a deterministic-first architecture: rules and validation logic are applied before and after any AI-assisted step, and outputs are never produced without passing validation.

• The mapping engine applies deterministic rules first. AI is only invoked for fields that rules cannot resolve.
• AI-assisted matches are assigned a confidence score and flagged for human review. They do not silently override rule-based results.
• All mapped fields pass through a validation layer before export. Required fields, type constraints and cross-field rules must pass before a clean export can be generated.
• The audit trail records every mapping decision, its source (rule-based, AI-assisted, or manual), and its confidence level, giving reviewers a full picture of how each field was resolved.
• Conservative defaults are used throughout: where a value is ambiguous or missing, the system flags it rather than filling it silently.

What validations are applied?

Helix applies layered validation at ingestion, after mapping, and again before each export is generated.

• Required field checks: Each output template has a defined set of required fields. Rows missing required fields are flagged and cannot be exported in that template until resolved.
• Type and format checks: Date fields, currency codes (ISO 4217), numeric fields and ISIN format (12-character alphanumeric) are validated against expected patterns.
• Consistency checks: Where fields should be consistent across a share class (e.g. domicile, fund group), contradictions between rows are surfaced.
• Cross-field rules: Template-specific logic is applied. For example, EMT requires consistent currency across investor profile fields; EPT requires performance identifiers.
• Row-level rejection messages: Every failed row returns a specific field-level reason, not a generic error. This allows targeted correction without re-processing the whole file.
• Validation results are included in the audit trail and visible in the review interface before export.

What does the audit trail include?

Every transformation is logged so that reviewers, compliance teams and operations managers can trace how each output was produced from its source.

• Source file reference: The original filename, upload timestamp and the user who uploaded it.
• Mapping decisions: Each field mapping is recorded with the method used (exact match, pattern match, AI-assisted, or manual override) and the confidence score assigned.
• Transformations applied: Any format conversions, type coercions or value normalisations are logged against the field and row they affected.
• Validation outcomes: Pass, fail or warning status per field and per row, with specific reasons for each outcome.
• User actions: Manual overrides, mapping confirmations and export events are attributed to the user who performed them.
• Timestamps: All events in the pipeline carry UTC timestamps for audit and compliance purposes.

How does Helix fit into our operating model?

Helix is designed to be adopted incrementally, running in parallel with your existing process before replacing it, rather than requiring a full cutover from day one.

• Parallel run: Start by uploading one fund's data and comparing Helix outputs against your existing templates. Validate before expanding scope.
• Incremental expansion: Once validated, onboard additional funds or templates. There is no hard limit on fund count or file volume.
• Role-based access: Different team members can be assigned different roles, for example upload access, review access and export approval.
• Sign-off workflow: Outputs can be reviewed and approved in the interface before being exported, supporting existing sign-off and governance processes.
• Helix does not require you to decommission existing systems as a precondition. It sits alongside your current stack and feeds it clean, validated data.
• For teams running manual EMT or EPT production in spreadsheets, Helix replaces that step without requiring changes to downstream processes or recipient systems.

Integrations and automation

Helix is designed to integrate with your existing technology stack via exports, API endpoints, and workflow and orchestration layers.

• File exports: All outputs are available as CSV or XLSX downloads, compatible with any downstream system that accepts structured files.
• API access: Helix exposes API endpoints for programmatic ingestion and export retrieval. API documentation is available on request.
• Workflow and orchestration layers: Helix is integration-ready for connection to workflow and orchestration tools that manage broader data pipelines and operational processes.
• Downstream systems: Clean, validated outputs are designed for onward delivery to distributor platforms, data warehouses, CRM systems and marketing tools.
• Automated or scheduled ingestion pipelines can be discussed as part of onboarding for teams with recurring data feeds.
• Speak to us about your specific stack and integration requirements: ludovic.milne@helix-ai.co.uk

Security and data handling

Helix processes data securely in transit and at rest, with access controls applied at user and role level.

• All data is transmitted over HTTPS. Data at rest is stored on secure cloud infrastructure.
• Access is controlled at user level. Role-based permissions limit what each user can upload, review or export.
• Uploaded files and generated exports can be deleted from the platform on request.
• Helix does not sell or share your data with third parties.
• For specific security questionnaires, compliance requirements or data processing agreements, contact us directly.
• Support: ludovic.milne@helix-ai.co.uk | Book a call: calendly.com/ludovic-milne-helix-ai

Commercials and onboarding

Helix is designed to be easy to start, easy to expand and easy to exit if it is not right for you.

• Free trial: Register at app.helix-ai.co.uk/register and upload your first file immediately. No credit card required to start.
• No onboarding fees: There is no implementation charge, setup fee or professional services requirement to get started.
• No minimums: No minimum fund count, minimum file volume or minimum contract value.
• Month-to-month: Pricing scales with usage. You are not locked into a long-term commitment.
• Typical pilot flow: (1) Register and upload a sample file. (2) Review mapped outputs. (3) Compare against your current templates. (4) Confirm scope and expand to your full fund range.
• For larger teams or firms with specific configuration requirements, we offer a guided onboarding session. Book via Calendly.

Frequently asked questions