Teenva AI & Digital Ventures builds reconciliation AI for finance, controllers, and accounting firms from Bangalore, India. Import bank statements and ledger exports—Teenva matches transactions by amount, date, reference, and parsed UPI/NEFT narrations, surfaces exceptions, and suggests one-to-many splits—then posts matches to Zoho Books or your ERP after reviewer sign-off.
Month-end bank rec in Excel breaks when volume grows or Razorpay/Stripe payouts batch hundreds of orders. Teenva automates high-confidence matches and queues exceptions with explainable reasons—closing the loop after invoice payments and expense reimbursements.
Build your AI solution with us
Reconciliation AI—bank vs books matching, payout reconciliation, and exception workflows.
Hub: AI solutions · Domain: Accounting
Reconciliation types we support
| Type | Bank / source | Match target |
|---|---|---|
| Bank vs cash book | HDFC, ICICI, SBI CSV/API (scoped) | GL cash account entries |
| Payment gateway | Razorpay/ Stripe settlement | Order receipts / AR |
| Vendor payments | Outgoing NEFT/UPI | AP bills |
| Payroll / reimbursements | Bulk transfer | Expense reports |
| Credit card (scoped) | Card feed | Expense lines |
| Intercompany (scoped) | Due to/from accounts | Elimination entries |
| Sub-ledger (scoped) | Customer deposits | CRM/ billing system |
Each type uses tuned rules + ML scoring—not one generic equality check.
How matching works
| Signal | Use |
|---|---|
| Amount | Exact or tolerance (fees, FX rounding) |
| Date window | T+1 settlement delays |
| Reference | UTR, cheque #, invoice # in narration |
| Narration parse | UPI/VPA, merchant prefix patterns (India) |
| Many-to-one | Single payout → multiple invoices |
| One-to-many | One customer payment → partial invoices |
| Recurring | Rent, SaaS subscriptions auto-link |
| Confidence score | Auto-match above threshold; else exception |
Optional LLM assist (scoped) explains why a fuzzy match was suggested—core match is rules + ML, not chat guessing.
Features we implement
Statement ingest
CSV, OFX, bank API (scoped), gateway reportsLedger import
Zoho/ERP export or API pullAuto-match engine
Batch run with audit logException inbox
Unmatched bank, unmatched book, amount mismatchSplit / merge UI
Accountant allocates one deposit to many linesRules library
“Razorpay settlement always fees 2%” configurableReconciliation report
PDF/Excel for sign-offPeriod lock
Prevent edits after close (scoped)Dashboard
% auto-matched, aging exceptions, time savedWebhook
Notify workflow when exceptions > threshold
Webhook notifies workflow automation when exceptions exceed threshold.
Architecture
Bank statement + ledger export (scheduled or manual)
Normalize amounts, dates, references (India UPI parsers)
Candidate generation + scoring model
Auto-post high confidence → ERP/Zoho
Exception queue → accountant review UI
Approved matches + adjustments exported
Metrics + training feedback from overrides
- Security — Financial data in your environment; role-based access
- Audit — Every match records rule version and user override
Use cases by business
| Business | Reconciliation pain |
|---|---|
| Ecommerce D2C | Razorpay settlements vs Shopify orders |
| SaaS | Stripe payouts vs subscription invoices |
| Agency / services | Client receipts vs project invoices |
| Retail (scoped) | POS deposits vs daily sales |
| NBFC / wallet (scoped) | Pool account vs user balances (high compliance) |
| Accounting CA firms | Multi-client bank rec at scale (scoped) |
Integrations
| System | Role |
|---|---|
| Zoho Books | Bank feeds via Zoho Books |
| Razorpay / Stripe | Settlement reports from Razorpay / Stripe |
| Custom ERP | GL via custom ERP |
| Transaction categorization | Enrich via transaction categorization AI (scoped) |
| API integration services | Bank aggregator via API integration services (scoped) |
99% match rate means nothing if the 1% is ₹10L unmapped.
We optimize auto-match + exception clarity—so accountants close books faster and trust the remaining items.
Reconciliation vs other accounting AI
| Product | Focus |
|---|---|
| Reconciliation AI | Bank/ledger/gateway matching |
| AI invoice processing | Capture vendor bills |
| Expense automation AI | Employee receipts |
| Accounting chatbot | How-to FAQ for finance users |
| Transaction categorization AI | PFM labels—not formal rec |
Why Teenva AI
India bank narrations
UPI, IMPS, NEFT reference patternsGateway + books
Razorpay/Stripe settlement logic we've integrated beforeAccountant UX
Split/merge and override—not black-box onlyMeasurable close
Hours on bank rec before/after pilotOps
Managed IT support when bank CSV formats change
Delivery process
Discovery
Accounts, banks, gateways, volumes, close calendar
Sample data
2–3 months statements + ledger (anonymized)
Baseline
Manual match rate and exception aging
Normalization
Parsers for narrations and gateway formats
Match engine v1
Rules + scoring; backtest on history
Exception UI
Accountant workflow UAT
ERP integration
Post matches to Zoho/ERP staging
Pilot
One account or entity for one close cycle
Production
All accounts; monitor auto-match % monthly
Frequently asked questions
Enhance or replace—we can push matches into Zoho or run standalone with export.
Scoped via aggregator or corporate banking API—CSV MVP common.
Date window rules in matcher; configurable per gateway.
High-confidence only; your threshold; exceptions always reviewed by default.
Scoped—FX amount tolerance and rate table.
No—scoring uses structured rules/ML; LLM optional for narration cleanup only (scoped).
Tenant isolation and per-client rules (scoped).
- Yes—match targets include AP payments and reimbursement batches.
10–12 weeks MVP (one bank + Zoho + CSV gateway report).
Match log export with user, timestamp, rule, before/after amounts.
Explore
Build your reconciliation automation
Build your AI solution with us—bank vs books matching, payout reconciliation, and exception workflows.
Build your AI solution with us
Ready to take your business to the next level?