Chai & AI
One AI concept a day, in a banker's language. Lessons run from tokens and embeddings through RAG, knowledge graphs, chains and agents, each one defined from zero and ending with a question for your desk.
Follow Chai & AI on LinkedIn →About
Most banks don't have an AI problem. They have a translation problem.
The technology works. The people who understand the trade desk and the people who understand the models are usually not the same people, and everything that stalls, stalls in between. That in-between is where I have spent my career.
I have spent 17+ years on one enduring challenge: how to make complex banking operations simpler, safer and more intelligent. The path ran from banking operations to product leadership and now AI-led transformation. Today I work at the intersection of enterprise AI, product and financial services, with deep expertise across trade finance, supply chain finance and payments.
The day job is sitting with operations, compliance, credit and treasury teams at global banks, turning real desk problems into working AI, and working AI into adoption the examiners can live with. Document scrutiny, compliance screening, and the unglamorous automation that actually moves a bank's numbers.
What I have learned in the trenches: AI succeeds in banking when the banker can narrate the use case. So I spend my mornings making that possible, in public.
The next generation of banking platforms will not be created by simply adding AI to legacy processes. It will come from redesigning work around intelligence, trust and human judgment.
Expertise
Where the work sits
Blue marks the trade side of the desk; violet marks the AI and product side. Most of the job is holding both at once.
Trade
Trade Finance, SCF & Payments
Letters of credit, guarantees, collections, supply chain finance and cross-border payments, from the operations floor to the product roadmap.
AI
AI in Document-Heavy Banking
Intelligent document processing, automated document examination, trade-based money laundering and compliance screening, agentic workflows with a human who signs.
Product
Product Strategy & Roadmaps
Turning customer discovery and field evidence into roadmap decisions, and separating what is positioning from what is configuration or integration.
Solutioning
Presales & Solution Design
Demonstrations, proofs of concept and bid responses that align to regulatory expectations and the way a bank actually operates.
Adoption
Customer Success & Value Realisation
Benchmarks before adoption, turnaround comparisons after it, and the handover from sales promise to delivered outcome.
Governance
Responsible AI in Regulated Work
Explainability, audit trails and the human who signs: AI governance a bank can show an auditor, not just a steering committee.
Journey
Operations floor to AI product
Three chapters, one line of work: trade finance, and the systems that run it. Each chapter taught a language the next one needed.
Trade
2009 — 2019 · THE DESK
Banking operations
A decade on trade finance desks at global and Indian banks, ending as head of a zonal trade operations unit.
- Import and export letters of credit, guarantees, collections, bill discounting and export finance
- Document scrutiny, audits, regulatory reporting and the five banking days
- User testing for a core trade platform migration, from the user's side
Trade × Product
2019 — 2025 · THE PLATFORM
Product leadership
Product innovation, development and presales at a banking software company, across trade finance, supply chain finance and payments.
- Roadmaps and prioritisation that took three banking products to market
- Hundreds of demonstrations and bid cycles with banks across five regions
- ISO 20022 readiness and SWIFT gpi capabilities across live implementations
AI
2025 — PRESENT · THE INTELLIGENCE
AI-led transformation
Leading solutioning, product delivery and customer success for AI-powered trade finance transformation with banks worldwide.
- From first demonstration to production: demo, pilot, adoption, renewal
- Document examination, compliance screening and workflow automation with a human in the loop
- Industry forums and daily public writing on AI in trade
MBA, International Business & Finance · PMP® · Executive education in data science, AI and digital leadership.
Writing
AI, in a banker's language
Three publications, one promise: no hype. Basics, cost, governance and the rails being laid under trade, explained the way a desk would explain them.
The Sight Draft
What actually moved in trade finance, supply chain finance and payments: tokenised bank money, shared ledgers, ISO 20022, the AI economy showing up in the trade book. Three reads and an honest counterweight.
Follow The Sight Draft on LinkedIn →At Sight
A monthly magazine at the intersection of trade and AI: the themes that need more than a post, from governance an auditor can live with to the economics of tokens and total cost of ownership.
Follow for the first edition →Simplifying Trade with Intelligence. ☕
Themes
What I keep coming back to
AI governance you can show an auditor
Explainability, audit trails and the human who signs. In regulated work the question is never "can AI do this?" but "which decisions may it make, which stay with rules, and where does accountability remain?"
The mandate matters more than the model.
The economics of AI
Tokens to total cost of ownership. Capability without cost literacy is a demonstration, not a deployment; a desk that lives by cut-offs should pick its queue on purpose.
Ask for the bill with the benchmark.
The digital rails under trade
ISO 20022, MLETR, structured data, tokenised bank money and shared ledgers. Trade finance is linked legs, and the rails arriving now settle them together or not at all.
Rails are ceremonial until the day they are infrastructure.
In person
Where you will find me
Industry forums
Trade finance conferences and banking industry seminars across India, Europe, the Gulf and beyond, speaking on AI in trade and trade-based money laundering.
Customer workshops
Executive conversations, product reviews and working sessions with operations, compliance, credit and treasury teams, translating desk problems into use cases a bank can adopt.
Say hello
A banker trying to make sense of AI, or a leader trying to make AI make sense to bankers
That is exactly the conversation I am here for. If you are exploring the future of AI in banking, transaction banking or enterprise product transformation, say hello. ☕