AI belongs on the operating agenda, with the chief executive accountable for turning experiments into changed work and measurable value. “The AI Rollout Belongs to the People Doing the Work” starts where that mandate becomes real: with managers and staff who know the workflow well enough to improve it.
A founder asked me what independence on a board should look like in practice. “What a Great Independent Board Member Actually Does” answers with a standard built around preparation, judgment, useful access, and the discipline to serve the company rather than the room.
On a payments call, I kept returning to the last manual step between approval and settlement. “Visa’s Agentic Pilot Puts Value-Added Services at the Supplier Portal” examines who earns the durable position when software begins entering portals, carrying credentials, and handling exceptions.
Execution is the test in every case: ownership must sit with people who understand the work, while trust comes from controls that survive contact with reality.
Share on LinkedIn: Visa’s Agentic Pilot Puts Value-Added Services at the Supplier Portal · What a Great Independent Board Member Actually Does · The AI Rollout Belongs to the People Doing the Work
The roundup
This week: Brookfield’s Anuj Ranjan says 1,200 AI initiatives across its portfolio have generated roughly $300M of EBITDA and ~$3B of enterprise value, and that AI must be CEO-driven, not IT-driven · Flagship Advisory says embedded-payments platforms need distinct acquisition and back-book migration motions rather than a passive referral model · Customers Bank says an AI-powered pilot cut select commercial-loan closings from 30–60 days to seven days · Nvidia released an open-weight model for document analysis that Ramp tuned for finance and Scale AI trained for merger due diligence · Fintech Takes guests argued that agentic finance needs performance guarantees and infrastructure accountability, not just terms and conditions.
The one thing
Brookfield’s Anuj Ranjan says 1,200 AI initiatives across its portfolio have generated roughly $300M of EBITDA and ~$3B of enterprise value, and that AI must be CEO-driven, not IT-driven.
Episode published 2026-10-20; Ranjan’s specific EBITDA-to-enterprise-value translation of AI initiatives at scale is a rare public data point from a $1.4T AUM operator.
My take: If AI must be CEO-driven and your CEO is still routing AI decisions through IT, you are not running an AI program — you are running a pilot, and the $3B enterprise value gap between those two things is now measurable.
Payments
Flagship Advisory says embedded-payments platforms need distinct acquisition and back-book migration motions rather than a passive referral model.
A newly published analysis opens a series on embedded-payments execution with specific go-to-market practices and a benchmark for annual back-book conversion.
My take: A vertical-software CEO has to run payments as its own commercial business, because an integrated product that the installed base never adopts creates no durable payment economics.
AI inside operating companies
Customers Bank says an AI-powered pilot cut select commercial-loan closings from 30–60 days to seven days.
The bank’s previously disclosed pilot result surfaced in current reporting as a measurable example of AI changing a finance workflow.
My take: A sponsor owning a commercial lender needs to underwrite AI against elapsed time, labor removed and credit controls preserved, because faster approvals are worthless if exceptions escape review.
Also: Nvidia released an open-weight model for document analysis that Ramp tuned for finance and Scale AI trained for merger due diligence. — Nvidia and Reflection AI introduced deployable open-weight models on Oct. 22, including versions designed for financial-document review and merger diligence.
Also: Fintech Takes guests argued that agentic finance needs performance guarantees and infrastructure accountability, not just terms and conditions. — The October 21 episode highlighted emerging companies using guarantees and performance pricing to address failures in agentic systems as payments and AI infrastructure converge.
I’d love to hear from you if you would like to discuss any of this further or have interest in working together. Reply to this email or write to er@ferncap.com. If a CFO or a deal partner you know would find this useful, forward it; they can sign up at ferncap.com/brief/.
Ernest
Fern Capital