Sanjeev Gupta

The problem

A group spanning 25+ legal entities — including Flipkart, Myntra, Cleartrip, Walmart India, F1 Info Solutions and Jeeves — with ~₹1,700 crore in monthly payments and a delivery network of hundreds of thousands of partners, all needing to be paid accurately and on time. At this scale, manual invoice handling doesn't just slow down; it breaks. Controls drift, reconciliation backlogs build, and a seven-day processing cycle quietly becomes the norm.

The approach

Rather than add headcount to absorb volume, I built the operation to remove manual effort. We deployed "Tap & Pay" — OCR-driven invoice automation — and stood up the "Wishmaster" payment ecosystem to handle self-invoicing, TDS, and UTR reconciliation for ~350,000 delivery partners. In parallel, we migrated to Oracle Fusion Cloud and rebuilt SOX controls around the new flow, so that scale and control reinforced each other instead of fighting.

The outcome

Invoice processing fell from seven days to two. SLA adherence reached 99% and first-time-right accuracy passed 92%. Rule-based auto-reconciliation now covers 95%+ of activity across 120+ bank accounts — and the close runs on Calendar-Day-1.

The problem

Airtel's finance ran the way most large telecoms did — decentralised, circle by circle, country by country. Twenty-three Indian circles and fourteen African markets each held their own processes, controls, and payment practices. It worked, but it couldn't scale, and it left the group exposed to inconsistency and risk that grew with every new market.

The approach

I helped build Airtel's global shared-services (GCC) model from the ground up — consolidating Accounts Payable out of 23 circles and migrating finance operations across 14 African countries into a single centralised centre in India. We implemented Oracle 10i and R12 under the ACE framework, built OBIEE dashboards for visibility, and pioneered Host-to-Host (H2H) ERP-to-bank integration for secure straight-through payments when that was still emerging in Indian enterprises. I also conceptualised and drove an automated expense-accrual engine — creating, actualising and reversing monthly accruals hands-free — which pulled the month-end close forward from Calendar-Day-4 to Calendar-Day-1.

The outcome

A fragmented, multi-country finance function became one controlled operation — standardised, auditable, and built to absorb new markets without rebuilding from scratch. The model became a template for how the group scaled finance internationally, run by teams of 40+.