Change the tires of a fast-moving car — Jay Sahal
What the modern CFO actually does at startup speed — agility, automation, and the discipline of keeping the business honest on numbers.
Jay Sahal joined Yellow.ai as CFO in December 2020 after fifteen years at Wipro — the last of them as CFO of the $1B+ Healthcare & Life Sciences business unit. Over that arc he watched Wipro grow from $800M to $10B. His metaphor for what the modern CFO does, especially in a company growing 100–300% a year, comes from that scale gradient: you're trying to change the tires of a fast-moving car, and you can't stop the car to do it.
The metaphor sounds romantic. In Jay’s hands it’s operational. Change the tires while moving means: an approval matrix that doesn’t stop the car, automation across the canonical paths that doesn’t ask the car to stop, and a reporting cadence fast enough to match the car’s speed. The work is unglamorous and largely invisible from outside the company — but it’s the difference between a business that scales and one that compounds errors. Jay’s career has been a slow accumulation of the specific moves that make this possible.
A career built watching one company twelve-x
Jay’s foundation isn’t a CFO seat — it’s fifteen years at Wipro across business units, last as CFO of a $1B+ business unit, watching the parent company twelve-x. That isn’t the same lens as watching it from the outside as an investor. Closing the books while a company multiplies by twelve installs an instinct for what scales and what doesn’t. The agility-to-process tradeoff isn’t theoretical for him. He’s seen what happens when the IT-services answer (predictable cadence, 5–15% growth) meets the startup question (100–300% growth, instant pivots) — and the answer he carries to Yellow.ai is a clear one: don’t apply the slow playbook to a fast company.
You are trying to change the tires of a fast-moving car. You can't stop the car and change the tires.
Agility yields to structure
The mechanism Jay uses is the approval matrix. Not a slowdown — a routing layer. “Each transaction is taken care not in a unique way but in a standard way.” When the field needs lightning response and the back office is still building the form, both sides lose. So Yellow.ai automated the four canonical loops in sequence: order-to-cash, procure-to-pay, hire-to-retire, record-to-report. Not all of them worked the first time. A billing-tool implementation got scrapped mid-build when it became clear it wasn’t going to land; the team swapped to a different tool and resumed. The lesson the v1 essay missed: failure mid-implementation is part of the discipline, not a deviation from it. You don’t get the path right the first time. You get it right by being willing to throw away the path that isn’t working.
Honesty as a CFO job description
Jay’s frame for the CFO’s value is unglamorous. The role is to be the one who keeps the business honest on numbers. He hears this from teams as feedback — “there’s an interesting CFO and he keeps us honest” — and it isn’t a compliance line. It means highlighting weaknesses and risks before achievements. Showing the variance the moment it’s spotted, not the quarter after. Doing “more dirty work than any other CXO,” in his words, because most functions naturally lead with what’s working — and someone has to lead with what isn’t, so the course-correction starts at the right moment.
I don't mince any word or hide any bad metrics. Before highlighting any good part, I tend to first highlight where things are wrong, where we need to improve.
The funding winter as stress test
The discipline showed up when cash got harder. Yellow.ai’s response to the 2022–2023 funding compression ran three moves in parallel. One, pitch Yellow’s own automation as the customer’s lever — turn the SaaS into the customer’s cost reduction or revenue lift. Two, restructure burn — Jay’s specific example: 15-person teams reorganized around what 5 people can do automated. Three, raise venture debt to extend runway without diluting at compressed multiples. The three moves are linked. None of them work without the operational discipline already in place: the approval matrix, the automated paths, the honest reporting. The funding winter didn’t create the discipline. It revealed it.
What an enterprise can give value to a customer will be linked to how they are operating internally — and CFOs will play a bigger role in driving that journey.
What to listen for
The full episode runs longer on the CEO-CFO partnership (Jay’s view: keep challenging outputs, keep the business honest on numbers, never bury bad metrics), what Yellow.ai did during the SVB week (no direct exposure, but a portfolio diversification anyway), and his calm under fire — “the panic button is really little far away from my hands.” Jay’s three-word descriptor is Hard Worker. Community Enthusiast. Value Creator. — borrowed from a beehive metaphor for what he’s trying to build at Yellow.ai. Listen at /podcast/ep-002-jay-sahal; for the longer conversation across the catalogue, see /topics/modern-finance-function.
Related questions
- How do CFOs operate when a company grows 100% a year?
- Jay Sahal's frame is mechanical. Don't try to slow the business to plan — you can't. Build the muscles that operate at speed: an approval matrix that handles 95% of transactions without exception, automation across order-to-cash and procure-to-pay so reporting cadence matches business cadence, and the discipline to call out numbers honestly when they slip — early, in the room, before they compound.
- What does a CFO mean by 'keeping the business honest on numbers'?
- Highlighting weaknesses and risks before achievements. Not waiting for the variance to surface in next quarter's review — putting it on the table the moment it's spotted. Jay's framing: most functions naturally lead with what's working. The CFO's job is to lead with what isn't, so course-correction starts at the right moment. Unglamorous work; asymmetric value — small early corrections beat large late ones.
- When should a startup put in formal financial processes like an approval matrix?
- Not when the team complains it's slowing them down — by then the cost has already compounded. The right trigger is exception-frequency: when transactions outside the standard pattern start dominating finance's workload, the automation is overdue. Jay's path at Yellow.ai: automate the four canonical loops — order-to-cash, procure-to-pay, hire-to-retire, record-to-report — before scale forces the work into manual exception-handling.
- How should a CFO respond to a funding winter?
- Three moves run in parallel. (1) Pitch the company's own product as the customer's automation lever — your service becomes their cost reduction or revenue lift. (2) Restructure burn — Jay's specific example at Yellow.ai: 15-person teams reorganized around what 5 people can do automated. (3) Raise venture debt to extend runway without diluting at compressed multiples. The discipline is doing all three at once, not sequentially.
Updates
- Rewrote in the v2 podcast-summary style — new editorial spine ('change the tires of a fast-moving car'), in-body quotes, Related-questions block.