Research Methodology

Our Approach

Intelligence that supports decisions — not replaces them

ipowork is a capital markets research and decision-support platform. We aggregate structured data from multiple reliable public and proprietary sources, apply AI-assisted analysis, and present intelligence in a format designed to help institutional professionals — merchant bankers, AIFs, lenders, and family offices — make more informed, more cautious decisions.

We do not provide investment advice. We do not guarantee outcomes. We earn only from distribution fees, never from issuer advisory mandates — which means our research has no structural incentive to favour any issuer. This independence is the foundation of everything we do.

Our core belief: Data-driven decisions are better than intuition-driven decisions. Better research reduces investment risk. And transparency about methodology builds more trust than broad claims of accuracy.

Data Sources

Where our intelligence comes from

ipowork aggregates data from the following primary and secondary sources. All sources are publicly available registries, regulatory databases, or licensed data providers.

🏛️

MCA21 / ROC Registry

Company master data, director information, charge registers, annual filings (AOC-4, MGT-7), incorporation documents.

Primary Registry

📊

Probe42

Structured financial data, legal history, charge sequences, EPFO records, GST filings, peer comparison, credit ratings, RPTs — accessed via licensed API.

Licensed API

⚖️

Court Records

Legal case history from High Courts, NCLT/NCLAT, DRT, ITAT, consumer forums, and district courts — aggregated via Probe42 and public e-court databases.

Public Records

🏦

CRISIL / ICRA / CARE / India Ratings

Credit rating actions, outlooks, instrument-level ratings, and rationale documents as publicly disclosed by rating agencies.

Rating Agencies

📋

SEBI Databases

SEBI debarment orders, settlement proceedings, consent orders, registered intermediary data, and AIF registration records.

Regulatory

🏭

EPFO / GST / RBI

Employee provident fund payment records, GST filing compliance, RBI enforcement actions, NBFC registration and licence data.

Statutory Bodies

📰

BSE / NSE Disclosures

SEBI Regulation 30 KMP disclosures, listed entity filings, board resolutions, shareholding patterns for listed group companies.

Exchange Filings

🔍

Reference Documents

Financial statements, charge documents, incorporation papers, and auditor reports downloaded from Probe42 or MCA for document-level analysis.

Document Layer

📌

Manual Verification

ED, CBI, SFIO enforcement actions; RBI PCA/penalty orders; NCLT insolvency petitions verified manually before each report.

Human-Verified

AIRA Score

How the AIRA Master Score is computed

The AIRA (AI-powered Readiness Assessment) score is a weighted composite of five intelligence pillars, each scored on a 0–100 scale. The overall AIRA score determines a company's distribution tier and readiness classification.

Pillar	Weight	Key Factors Assessed	Primary Data Sources
⚖️ Legal Intelligence	25%	Wilful defaulter status, SEBI/RBI debarment, case volume and severity (NCLT, DRT, High Courts), ED/CBI/SFIO actions, BIFR/CDR history	Probe42 legal history, SEBI orders database, MCA compliance data, manual regulatory verification
🔬 Forensic Intelligence	20%	Audit opinion quality (qualified/unqualified), going concern flags, RPT exposure as % of revenue, EPFO/GST compliance, auditor changes	AOC-4 financial filings, auditor's comments, Probe42 RPT data, EPFO payment records, GST compliance records
👤 Director & Promoter Intelligence	20%	Board composition, Independent Director count, women directors, DIN deactivation flags, promoter holding stability, KMP tenure	MCA director records, DIN database, Probe42 director association history, shareholding pattern
📊 Financial Intelligence	25%	Revenue CAGR, EBITDA margin vs peers, ROE, D/E ratio, ICR, cash flow quality, peer comparison benchmarking, export revenue	AOC-4 financial statements (P&L, Balance Sheet, Ratios), Probe42 peer comparison, credit rating rationale
🏦 Credit Intelligence	10%	Credit rating grade and outlook, interest coverage ratio, security perfection, charge register quality, lender diversity	CRISIL/ICRA/CARE rating disclosures, Probe42 charge register, open/satisfied charge sequences

Derived readiness scores are computed from weighted combinations of pillar scores to indicate suitability for specific capital market pathways:

IPO Readiness = Legal×25% + Financial×35% + Director×25% + Forensic×15%
Debt Readiness = Credit×35% + Financial×35% + Legal×20% + Forensic×10%
PE/VC Readiness = Financial×30% + Director×30% + Forensic×25% + Legal×15%
Governance Score = Director×40% + Legal×30% + Forensic×30%

Analysis Process

How each report is generated

1

Data Collection

Structured data is retrieved via the Probe42 API (comprehensive-details endpoint) for the target company. Reference documents — financial statements, charge documents, incorporation papers — are downloaded from the Probe42 dashboard.

2

Manual Regulatory Verification

ipowork analysts manually verify ED, CBI, SFIO, SEBI, RBI, NCLT, DRT, and income tax enforcement actions. These are not available via any single API and require cross-referencing multiple regulatory databases.

3

AI-Assisted Document Analysis

Uploaded PDF documents are analysed using Claude (Anthropic's AI model) to extract audit opinions, RPT disclosures, financial parameters, charge details, and legal filings. AI reads multiple documents simultaneously and identifies material disclosures.

4

Pillar Scoring

Each of the five pillars is scored 0–100 using a deterministic methodology — base scores with defined deductions and additions based on specific criteria. Scores are computed by the analyst with AI assistance and reviewed for consistency.

5

Report Generation & Review

A structured AIRA Master Intelligence Report is generated with pillar summaries, risk flags, analyst recommendations, and a verdict. Reports are reviewed by ipowork's analysis team before publication.

6

Delivery

Reports are delivered via a secure, branded online portal with a unique URL. Co-branded or white-label delivery is available for merchant banker and institutional partners. PDF download is available on the report viewer.

AI Methodology

How artificial intelligence is used

ipowork uses AI as an analytical assistant — not as an autonomous decision-maker. All AI-generated outputs are reviewed by human analysts before being included in any report or intelligence product.

Claude (Anthropic) — Our AI Engine

ipowork uses Claude, developed by Anthropic, for document intelligence and research assistance. Claude reads uploaded financial statements, legal filings, charge documents, and auditor reports to extract structured insights across our five intelligence pillars.

AI is used for: document reading and extraction, pattern recognition across large datasets, flag identification, and report drafting. AI is not used for: final scoring decisions, regulatory status verification, or publication without human review.

Document Reading Data Extraction Flag Identification Report Drafting Human Reviewed

Data Currency

How frequently data is updated

Different data elements have different update cycles depending on the underlying source. Users should note the data vintage when reviewing any ipowork report.

Data Type	Source	Typical Refresh
Company master data (status, registered address, CIN)	MCA21 via Probe42	Daily
Charge register (open/satisfied charges)	MCA21 via Probe42	Daily
Legal history (case filings, status)	Probe42 / eCourts	Weekly
Financial statements (AOC-4, P&L, Balance Sheet)	MCA21 filings via Probe42	Annual (post-filing)
Credit ratings	CRISIL / ICRA / CARE disclosures	On rating action
GST compliance	GSTN via Probe42	Monthly
EPFO payment records	EPFO via Probe42	Monthly
Regulatory enforcement actions (ED, CBI, SEBI, RBI)	Manual verification	Per Report
Director DIN status	MCA21 via Probe42	Daily
Shareholding pattern	MCA21 annual filings	Annual (post-filing)

Limitations

What ipowork cannot tell you

Transparency about limitations is as important as transparency about methodology. The following constraints apply to all ipowork intelligence products.

Source Dependency

ipowork's accuracy is dependent on the accuracy of underlying sources — MCA, Probe42, rating agencies, and court databases. Errors, delays, or omissions in source data will be reflected in ipowork outputs.

No Private Information

ipowork cannot access private contracts, internal management communications, undisclosed related party transactions, off-balance-sheet liabilities, or any information not in public registries or filings.

Financial Data Lag

Financial statements are available after MCA filing — typically 6–18 months after the financial year end. ipowork reports reflect the most recently filed data, which may not represent current company performance.

No Forward-Looking Guarantees

AIRA scores reflect historical and current data. They do not predict future performance, future regulatory actions, or future market conditions. A high AIRA score does not guarantee investment returns.

Not Investment Advice

ipowork provides research and intelligence, not investment advice or recommendations to buy, sell, or hold any security. Users must conduct independent due diligence and consult registered advisors.

AI Limitations

AI-assisted analysis may occasionally misinterpret complex legal language, unusual financial structures, or ambiguous disclosures. All AI outputs are reviewed by human analysts, but errors may occur.