AI for the CFO
A Chief Financial Officer can already make AI produce a forecast. The hard part is the one nobody else owns: defending the number in front of a board when the future will not hold still. The Havruta Methodology (formerly the Think Partner Methodology) turns AI on that decision, so the machine reasons from your real drivers, attacks the soft assumptions, and helps you stand behind the forecast, instead of handing you a confident figure you cannot account for.
The finance question is almost never "can AI forecast this?". It can, quickly, which is exactly why the output is not the point. The decision a CFO owns is the defence: turning drivers, cohorts and scenarios into a number the board will trust, and being able to say why it holds and where it breaks. Gildoni installs the Havruta Methodology into how a CFO reasons with AI on that forecast, so the machine reasons from your real data rather than a plausible model, stresses the assumptions you are least sure of, and leaves you with a defensible number, not a confident one. This is not an FP&A tool. It is the reasoning discipline behind the forecast.
Producing a forecast is easy now. Standing behind it is the job
Every treasury and FP&A vendor will now generate a cash-flow forecast in seconds, and the forecasts are slick. Yet the CFO's hardest moment has not changed, because it was never the production. It is the board meeting where someone asks why the number is what it is, what happens to runway if the top two assumptions are wrong, and whether you would bet the hiring plan on it. A forecast you cannot account for, however confident, is a liability the moment it is questioned. The work is the defence, not the draft.
That defence is reasoning, not computation. It is knowing which drivers actually move the number, which assumptions are load-bearing, how the scenarios fan out, and where the forecast is fragile to a change you can already see coming. It is sharpest where cash is the live question: a SaaS or high-growth business managing runway, a PE-backed company defending a plan to its sponsor, a finance team setting a number the whole budget hangs on. None of that can be answered by a model that does not know your real drivers and cohorts. That is exactly what the methodology puts the machine to work on.
A forecast.
A forecast you can defend.
Inputs and a model template.
Your real drivers and cohort data.
Where the number is fragile.
The assumption that breaks it first.
What the methodology installs for a CFO
The Havruta Methodology changes the default. Instead of a machine that hands over a number and a confidence band, it installs the discipline that makes AI argue the forecast before it endorses it, which for a CFO maps directly onto defending the plan.
The Flip turns the machine on the forecast. Rather than confirming the model you built, it attacks it: which assumption is this number most sensitive to, what does the downside case actually look like, where would a board member press first, what are you treating as given that is not. You make it stress-test the forecast before the board does.
Ground Truth keeps it honest. A forecast built on an AI that has invented your churn, your sales cycle or your cohort behaviour is worse than a spreadsheet, because it is confident and wrong. The methodology insists the machine reason from your verified drivers and history, not a generic SaaS curve, so the number is about your business, not a plausible average of someone else's.
And the output is what the method calls a defensible gap: an arguable, evidence-backed account of the number and its risks, not a single figure presented as certainty. Reasoning with AI as a partner compresses the distance from raw drivers to a forecast you can stand behind, without surrendering the judgement to the machine. The fuller account of how all of this works is on the methodology page.
What this is not
This is not a forecasting tool, and it is not for your stack. It is not an FP&A platform, a treasury system, a planning model, or another product to evaluate. It is not AI training, and it is not generic AI literacy. It changes how you, the CFO, reason with AI on the decision you own: the forecast you commit and defend, and the calls that hang on it. The tooling is a separate, crowded problem. This is about the judgement.
- An FP&A tool
- A treasury platform
- A forecasting model
- A planning system
- Another tool to evaluate
- AI training
- Generic AI literacy
This is about the judgement.
Where a CFO starts
The methodology is installed along a ladder, and a CFO enters at the rung that fits. Most begin with the Eye-Opener Workshop, a half-day in which a finance leadership team sees the shift on its own real forecast. A CFO who wants to embed the discipline personally goes deeper through the Executive 1-1 Coaching Programme; a finance leadership group builds the rhythm through the Havruta programme; and a single high-stakes question, the board forecast, the runway plan, the financing decision, can be worked end to end through Advisory Havruta. A Strategic Briefing is how to decide which fits.
- Most begin here
Eye-Opener Workshop
A half-day in which a finance leadership team sees the shift on its own real forecast.
- For the individual CFO
Executive 1-1 Coaching Programme
The deeper, individual rung for the leader who owns the number.
- For the finance leadership group
The Havruta programme
An ongoing rhythm that embeds the practice across the finance team.
- For one high-stakes question
Advisory Havruta
The board forecast, the runway plan, the financing decision, worked until it is answered.
Frequently asked questions
How should a CFO use AI for cash flow forecasting?
Use it to defend the forecast, not just to produce one. The value is not a faster number; it is a machine that reasons from your real drivers and cohorts, stresses the assumptions the number is most sensitive to, and shows you where runway breaks under a downside case, so you can stand behind the figure in front of a board. The model is an input; the defence is yours. The Havruta Methodology installs that as a repeatable discipline, so AI sharpens the forecast judgement instead of just generating a projection.
What is cash flow forecasting?
Cash flow forecasting estimates the cash moving into and out of a business over a future period, to anticipate runway, liquidity and financing needs. AI has made the projection itself cheap and fast, which shifts the value to the reasoning around it: which drivers move the number, how confident each assumption is, and what the forecast does under stress. A projection is arithmetic; a forecast you can defend is a judgement.
Can AI do financial forecasting?
Yes, and that is the least interesting thing about it. AI can build and update a forecast faster than any analyst, which is precisely why the work moves to the part it cannot do for you: deciding which assumptions to trust, defending the number to the board, and owning the consequences. Used well, AI is the sharpest stress-tester of your forecast in the room, not the one that signs it.
How is this different from an FP&A or treasury tool?
An FP&A or treasury tool generates and maintains the forecast. This changes how you reason with AI to defend it, anchored in your real drivers rather than a model template. The tooling market is crowded and capable; this is a different problem, your own financial judgement under scrutiny, which no tool addresses.
Who is it for, exactly?
CFOs, finance directors, and the FP&A and finance leadership teams around them who own the number the board trusts. It is sharpest in SaaS, high-growth and PE-backed businesses where cash runway is the live question, and it suits leaders who are past asking whether AI can forecast and are now deciding how to stand behind the forecast it produces.
Where should we start?
With the Eye-Opener Workshop. It is the gateway: a half-day, built around your own real forecast, where the difference between instructing a machine and thinking with one becomes obvious. A Strategic Briefing is the fastest way to map the right entry point for your finance leadership.