TransactIQ
Bank statement analyzer API by Terra Insight Pvt. Ltd.
Underwriting infrastructure for Indian NBFCs and digital lenders. Benchmarked on PSU and co-operative bank formats, and engineered for the credit signals a lender actually uses — not a dashboard the underwriter has to re-interpret.
Three prongs, one frame
Most NBFC CTOs evaluating bank statement analysis are choosing between the cost of running BSA at scale, the accuracy of what comes out, and the depth of the signals they can act on. TransactIQ is designed around all three, not one.
Architected for materially lower unit economics
TransactIQ is designed so that the cost of running BSA at scale is not the binding constraint on how many loans a lender can underwrite. Self-hosted deployment on your infrastructure, efficient inference, and batch-friendly pipelines. Lenders running meaningful BSA volume see unit economics substantially below incumbent managed tiers.
Accuracy on the statements where others fail
Every vendor works on a clean HDFC PDF. The real underwriting bar is degraded statements — PSU bank dot-matrix scans, Karnataka State Co-operative statements, Dhanlaxmi password-protected PDFs, district central co-op formats. TransactIQ is benchmarked quarterly on exactly those inputs because that is where NBFC credit decisions actually break.
40+ engineered features, not line-item extraction
Line-item extraction is table stakes. What an NBFC credit team needs is bounce prediction, salary consistency scoring, round-tripping detection, EMI obligation aggregation across lenders, and synthetic P&L/balance-sheet construction for MSME borrowers. TransactIQ produces those as API outputs — not as a dashboard the underwriter has to re-interpret.
Synthetic financial statements for MSME borrowers
Most Indian MSMEs have no audited financials. Incumbent BSA vendors stop at line-item extraction. TransactIQ goes four layers deeper — constructing synthetic P&L, balance sheet, and cash-flow views directly from bank activity, so credit teams can underwrite the ₹65-trillion MSME credit gap without waiting for statutory audits.
Personal vs business separation
Before a P&L is even possible, you have to know which transactions belong to the business and which are the proprietor's personal spending. TransactIQ uses counterparty intelligence and behavioural patterns to split a single bank statement into two implicit ledgers.
Synthetic P&L
Revenue segmented by channel, direct and indirect costs inferred from narration-level categorisation, and an EBITDA figure cross-validated against GST filings where available. Usable as an income proxy for MSME underwriting even when books are informal.
Synthetic balance sheet
Current assets and liabilities, working capital position, visible borrowings (EMI outflows, overdraft activity), and an indicative net-worth figure. Not an auditor-grade balance sheet — a decisioning-grade one, which is what credit actually needs.
Synthetic cash flow
Operating, investing, and financing cash flows inferred directly from bank activity — the structural complement to the P&L and balance sheet. The three-statement view that no other Indian BSA vendor produces today.
Core capabilities
What TransactIQ does on every bank statement it processes, and what a CTO evaluating it should expect from the API.
OCR on degraded inputs
Built for dot-matrix PSU scans, password-protected PDFs, multi-column layouts, and cooperative bank formats where incumbent OCR degrades below useable accuracy.
Transaction categorisation
UPI, NEFT, RTGS, IMPS, cheque, cash deposit — every transaction typed by rail and counterparty class. Narration parsing is India-specific, not ported from US or European models.
Credit signal engineering
Bounce prediction, salary consistency scoring, EMI obligation aggregation across lenders, cash-flow volatility, round-tripping detection, mule-account patterns — as direct API outputs.
API-first delivery
Sync, async, and webhook endpoints. Batch processing up to 10,000+ statements per day. Idempotent, versioned, with a sandbox the CTO can hit before talking to sales.
Data residency and audit
AWS Mumbai region, RBI IT governance–aligned data localisation, DPDP Act 2023 consent architecture, SAR-ready audit trail of every processed statement.
Deterministic outputs
Every signal is rule-traceable. No black-box model scores the credit team cannot explain to a regulator. Fair-lending defensible, audit-defensible, model-risk defensible.
Bank coverage
TransactIQ covers every statement format an Indian NBFC is likely to receive — including the cooperative and PSU formats where incumbent vendors routinely degrade to manual re-entry.
HDFC · ICICI · Axis · Kotak · IndusInd · Yes · IDFC First · RBL
SBI · Canara · PNB · BoB · Union · Indian · BOM · UCO · Central Bank
District central co-ops · Urban co-ops · State co-ops · Karnataka State · TJSB · Saraswat
Ujjivan SFB · Equitas SFB · AU SFB · ESAF SFB · Airtel Payments Bank · Paytm Payments Bank
Deployment tiers
Three deployment shapes, matched to lender stage and isolation requirements. Data residency on AWS Mumbai in every case.
TransactIQ runs inside your VPC on your infrastructure — AWS, Azure, or GCP India regions. You own the data plane. Lowest unit economics, highest data-residency control.
Multi-tenant managed tier on AWS Mumbai. Fully operated by Terra Insight. No infrastructure effort for your team. Ideal for smaller NBFCs and digital lenders scaling up.
Dedicated tenant on AWS Mumbai with isolated compute and storage. Single-tenant data plane with managed operations. For enterprise NBFCs with stricter isolation requirements.
TransactIQ in depth
Every layer of the product has its own page — explore whichever is relevant to your evaluation.
Join the TransactIQ early-access list
TransactIQ is in structured rollout through 2026. Early-access lenders get API sandbox credentials, accuracy-benchmark reports, and direct engineering partnership through the integration window.