Chartered Accountant | Finance Leader | Automation Enthusiast
Applying AI and automation to real finance workflows with a regulator's spine. Where reliability, auditability, and human oversight aren't optional—they're foundational.
Three principles that guide everything I build
AI should assist decisions, not replace accountability. Every system must have explicit approval gates and clear boundaries for what AI can and cannot do autonomously.
Ambiguity is the enemy of trust. Every system must clearly define its scope, authority, and escalation paths—ensuring humans remain the final decision-makers.
Systems should scale across use cases, not just demos. I build architectures that adapt to new contexts without requiring complete rebuilds.
AI-assisted workflows with built-in accountability
An AI advisory assistant for finance, tax, and automation queries. Uses an internal knowledge base with retrieval-augmented generation, real-time web search only when required, and strict guardrails to escalate case-specific matters to human consultation.
A centralized AI assistant that interprets business intent and orchestrates emails, meetings, spreadsheets, and contacts. Designed with intent-based authorization and explicit confirmation for sensitive actions—demonstrating how AI can be a helpful orchestrator without being a decision-maker.
AI-assisted extraction and classification of expense documents with validation rules, suggested accounting entries, and mandatory human approval before final posting. Showcased at the ICAI AI Hackathon to demonstrate how AI can assist (not replace) accountants in routine workflows.
End-to-end automation for managing book borrowing, returns, and extensions for an internal book club. Includes availability checks, status updates, notifications, and a complete audit trail of book movement. Built using Microsoft Power Platform—demonstrating solid workflow automation principles.
Automates the earning, approval, and utilization of compensatory off hours with SLA-based rules, auto-approval on inaction, and a complete audit trail for HR and management. Demonstrates policy-as-code principles—translating HR policies into automated workflows.
Automates scheduled LinkedIn posting using a topic calendar stored in Google Sheets. Prevents duplicate posts, maintains posting continuity, and removes the need for daily manual content execution.
Building solutions from 2017 onwards
Built during GST transition with user authentication, automated invoice numbering, organized storage, email notifications, and comprehensive activity logging with fraud prevention controls.
Automated GST return filing for 17 registrations. Generated JSON files for bulk upload, reducing 3 man-days to 3 man-hours while eliminating errors.
Emergency solution for 20,000+ monthly transactions. Developed nested formulas to automatically map credit notes to original invoices, filing 10 months of accumulated returns.
Automated verification for monthly payroll. Power Query-based workbook performs all validations automatically and outputs only items that fail checks.
Redesigned MIS templates from scratch for board and MD reporting. Collaborated on Tableau dashboards for automated weekly reporting.
Comprehensive tracking for audit tasks with dynamic visualizations (speedometer, dumbbell charts) showing completion status, delays, and timelines.
Six years from audit to automation
I am a Chartered Accountant (CA) and ACCA professional with six years of experience in statutory reporting, financial consolidation, and audit across global markets. My career spans from article training at a CA firm to managerial roles in listed companies and shared services, where I've coordinated with Big 4 auditors, managed multi-jurisdictional statutory reporting, consolidated financial reporting & SEBI reportings, and led system transformations affecting 40+ global entities.
My journey with automation started early—during article training in 2017, I built VBA-based billing systems during the GST transition and automated complex GST return filings. What began as solving immediate problems evolved into a consistent pattern: identifying manual fatigue and building solutions that maintain control while eliminating repetitive work.
Over the years, I've built everything from Excel-based verification systems to AI-powered assistants, always with the same principle: technology should assist, not replace accountability. This approach comes from working in environments where consequences matter—audits, statutory compliance, board-level reporting—and where errors have real implications.
I've trained over 150 finance professionals in advanced Excel, Power Query, Power Automate, and Power BI, and an active trainer and mentor for graduates and professionals, helping teams adopt technology thoughtfully. I'm currently actively exploring AI applications, experimenting with how they can be deployed safely in finance operations while respecting the need for traceability, human oversight, and clear boundaries.
A builder with a regulator's spine
"You automate human fatigue. Not creativity. Not judgment. Fatigue."
I prefer systems that explain themselves, fail predictably, and leave a trail. I naturally gravitate toward approvals instead of assumptions, logs instead of memory, rules instead of vibes.
I think in layers: input, interpretation, decision, execution, review. AI interests me as a middle layer—something that can assist, route, and suggest, but should never be the final authority.
When something goes wrong, my first reaction isn't to ask who caused it but why the system allowed it. This mindset comes from years of working with auditors, regulatory authorities, and CFO-level reporting.