The thesis: finance has always worked one way — people execute, systems record. That division of labor is now inverting: AI agents run the processes end to end — the close, payments, reconciliations, compliance — under the enterprise’s policies and controls, routing only exceptions to people. The result is finance gone lights out: self-running, exception-based, always on — through month-end, quarter-end, and Day-1 events.
Every paper on this site argues that thesis into one part of the business — with a working model you can run on your own numbers, and a live benchmark to measure your operation against.
One thesis — defined, layered, and measured here — argued across six parts of the business.
The close, treasury, the economics of the machines, and the org chart that survives them.
The papers ↓The systems of record agents operate, and the data foundation that sets the ceiling.
The papers ↓The control plane, the regulatory dividend, and financial-crime operations rebuilt.
The papers ↓The end of the ticket: shared services re-chartered around outcomes.
The papers ↓Order-to-cash run autonomously — and paid out twice.
The papers ↓Post-trade on the market’s clock, and the digital-asset back office end-to-end.
The papers ↓Every paper is interactive: the argument, the architecture, and a model you can run on your own numbers rather than a claim you have to take on faith.
The monthly close is a batch ritual on a streaming business. What it takes to make closing a standing condition instead of a campaign.
Read the paper →Every treasury has two policies: the approved one and the operating one. Executable policy makes them the same document for the first time.
Read the paper →Autonomous operations create a new cost line — and finance is both payer and operator. FinOps keeps the economics of autonomy honest.
Read the paper →Pilots don’t die of weak models; they die of unchanged operating models. The four roles that remain — and the career ladder that must be rebuilt.
Read the paper →Thirty years of ERP taught the enterprise to remember. Clean core and policy-as-configuration are turning it into a platform agents can operate.
Read the paper →Why agents amplify defects, the autonomy-ceiling model, the minimum viable foundation in four bounded moves, and the standing patrol that should ship before any transactional agent.
Read the paper →Autonomy that cannot be assured is a finding waiting for fieldwork. The control plane that makes agents attestable — and the flip from sampling to census.
Read the paper →T+1, MiCA, DORA, AI regimes, e-invoicing: one specification in five rulebooks. The dividend calculator prices what a shared control plane returns against point solutions that die with their mandate.
Read the paper →Financial-crime operations manufacture documentation: 95% false positives triaged by arrival date. The alert-factory model shows what changes when agents investigate and humans judge.
Read the paper →Why the operating model — not regulation, custody, or liquidity — is the binding constraint on institutional digital assets. With live market data, the five-layer operating model, an ROI model, and the Lights Out Maturity Index.
Read the paper →Microsecond execution, days-long operations. What the quant desk’s autonomy disciplines teach post-trade — and the collateral opportunity nobody staffs.
Read the paper →Every paper is opinion until it meets data. The Lights Out Maturity Index — six questions, two minutes, plain-language answers — measures where finance operations actually sit on the path from spreadsheets-and-heroics to autonomous. Anonymous responses build the benchmark this publication reports on: maturity by domain, by industry, by region.
The Lights Out Maturity Index: six questions, two minutes, no scales to interpret. Your anonymous result joins the inaugural Lights Out Finance Survey — the benchmark this publication reports on.
Take the Index Browse all papersLights Out Finance is written and edited by Adil Bahir — a Big 4 partner in CFO Advisory & Finance Transformation with more than twenty years across the Americas, EMEA, and APAC. His delivery record spans the Day-1 finance model for a $9B TMT divestiture, a $2.1B asset-management carve-out across 24 countries, reconciliation automation above 90% touchless across 37 countries, and $50M–$100M+ ERP transformation programs — the operational benchmarks referenced throughout these pages.
He holds a Doctor of Engineering (DEng.) in Artificial Intelligence from George Washington University, an MBA (Finance) from Cornell University, a Master in Financial Engineering from Queen’s University’s Smith School of Business, and an MEng/MBA from École des Ponts ParisTech; professional credentials include US CPA, CGMA, FRM, CQF, CTP, and CDAA. His work spans the full breadth of financial technology — ERP and core-banking modernization, treasury and capital markets platforms, quantitative finance, FinOps, and enterprise AI — with one thread throughout: taking finance functions to autonomous, exception-based operation.
Read the full masthead — the thesis, the framework, and the editorial principles →
Views expressed are the author’s own and do not represent any employer or affiliated organization.
New papers land in your inbox — enter your work email below. Benchmark respondents who leave an email receive the survey report directly when it publishes. RSS available for readers who prefer it.