You have ten minutes on the audit committee agenda, one slide, and two people in the room who will read it in opposite directions. The CFO will look for the return: how much faster, how much cheaper. The chair of the audit committee, with the assurance partner sitting just behind, will look for the exposure: what does putting AI near a disclosed number do to the risk that the number is wrong. If your slide answers only the CFO, the chair hears an unmanaged risk and the item stalls. If it answers only the chair, the CFO hears cost with no return and the item dies. The entire craft of taking reporting AI to the C-suite is fitting both answers into one honest view, so that the efficiency gained and the assurance risk reduced sit side by side and neither is hidden behind the other. This lesson is how you build that view and present it without overclaiming a single number.

The distinction that matters here is not how to compute the business case from scratch. It is how to present it to the two hardest readers in the building at once. A CFO and a board do not want your working; they want the shape of the decision and the honesty of the person presenting it. So this lesson is about the dual-axis story as one view: what goes on the axes, how you show each without pretending to a precision you do not have, and the sentences that keep the whole thing credible when the assurance partner is listening for the first overclaim.

There is a reason this is a strategist skill and not a spreadsheet skill. The numbers can be perfect and the presentation can still fail, because the C-suite is not evaluating the numbers in isolation; it is evaluating whether the person bringing them understands the risk the company is actually carrying. An audit committee that has lived through a restatement, or watched a peer live through one, reads a reporting AI proposal as a question about judgement before it is a question about return. Show them a proposal that treats assurance risk as an afterthought and you have told them, without meaning to, that you do not yet understand the environment you are operating in. Show them a proposal that leads with the risk, prices it honestly, and pairs it with a return, and you have told them the opposite. The one view is a demonstration, not just a request.

Two readers, one view

Start by naming who is in the room, because the audience defines the slide. A CFO reads for return and thinks in cost, cycle, and payback. A board, and especially an audit committee, reads for exposure and thinks in the probability and consequence of the disclosure being wrong. For the companies still inside CSRD after the Omnibus, which are the largest undertakings, a failed sustainability disclosure is not a line-item correction; it is a board-level event, with audit-committee scrutiny, possible restatement, the cost of re-assurance, and a reputational hit that can reach financing terms. That is why the risk reader in the room is not a formality. External assurance is now the norm rather than the exception for large global companies, and the trajectory runs from limited toward reasonable assurance, the more demanding standard. The person who cares whether the number survives the assurer has real power over whether your program proceeds.

A single-axis story fails both readers at once. Sell only efficiency and the audit committee hears a program that makes the company faster at producing numbers nobody has confirmed are defensible, which sounds like accelerating toward a cliff. Sell only risk reduction and the CFO hears spending with no return, a compliance tax. The dual-axis view is not a presentation trick; it is the only frame that speaks to both readers truthfully, because a genuine reporting AI win really does move both. The discipline that makes an AI output traceable, every figure pointing back to a source and a documented method, is the same discipline that makes the next cycle faster, because the evidence is already assembled. Efficiency and assurance are not a trade-off to be balanced. In a well-built program they are two readings of the same underlying improvement, and your one view should show them as such.

To a CFO you are selling return; to the audit committee you are selling exposure reduced. Put both on one slide, or one of the two readers in the room hears only half a story, and the half they hear is the one that stops the program.

The efficiency axis, told without inflation

The efficiency axis is what the CFO came for, and it is the axis most likely to be overclaimed, because the numbers are concrete and a vendor has already handed you a flattering percentage. The strategist's job in the C-suite is to present efficiency in a form that survives the CFO's own scrutiny, which means showing where the number came from and what has already been subtracted.

Present efficiency as two figures a CFO recognises: a change in cycle time and a change in cost-to-report. Cycle time is elapsed weeks from a frozen start to a frozen end, and you say the boundaries are frozen, because a CFO knows that a faster cycle achieved by moving the finish line is not a saving. Cost-to-report is built from your own analyst hours at a fully-loaded rate, not from a vendor's headline percentage, and you say so plainly: the time-saving in this slide is the one measured on our own data in a proof of concept, not the one in the brochure. Then you show the two subtractions most presentations hide, because showing them is what makes the CFO trust the number. First, the oversight added back: AI shifts work from doing to reviewing rather than deleting it, so someone still checks the figures, and those hours are real and are already netted out. Second, the platform cost: licence, implementation amortised over its life, and integration, subtracted before you call anything a saving. A net efficiency number that has visibly survived those subtractions is worth more in the room than a gross number twice its size, because the CFO has seen you resist the inflation before they had to point it out.

The risk axis, priced as a range not a decimal

The risk axis is what the audit committee came for, and it is the axis most likely to be either omitted or faked. Omitted, because it is hard to quantify and does not appear in vendor decks. Faked, because someone reaches for a single confident number to match the efficiency figures, and a false decimal on a risk estimate is the fastest way to lose the assurance partner in the room. The move that works is to price the risk axis honestly as expected value and as a range, and to say out loud that it is a judgement.

Frame the risk axis as the reduction in expected cost of a disclosure failure. It has two components, both of which a board understands from every other risk they price. The consequence: what a restatement of the sustainability disclosure plus the required re-assurance and management time would cost, before reputational and financing effects, presented as a range because it genuinely is one. The probability change: how much a cleaner, reconstructable basis of preparation lowers the annual chance of such an event, presented as a range of percentage points and labelled explicitly as informed judgement rather than measurement. Multiply the two and you have an expected avoided cost, shown as a range. For the largest in-scope companies the consequence is so large that even a modest probability reduction produces an expected avoided cost that can rival or exceed the entire efficiency saving. That is the fact that turns the risk axis from a soft compliance argument into the harder half of the financial case, and it is why the audit committee's concern is also, quietly, a CFO argument.

Tie the probability reduction to named mechanisms so it is not hand-waving. The basis of preparation is reconstructable on demand, so an audit request does not trigger a scramble. Every figure traces to a source, so findings from unsupported numbers do not arise. Methods are documented and consistent year over year, so the prior-period comparison holds. Provenance is tagged and primary data is labelled distinctly from estimates, so nothing is laundered. Each mechanism is a concrete reason the probability of a finding falls, and naming them lets the audit committee interrogate the claim instead of taking it on faith. A risk reduction the chair can interrogate is far more persuasive than one presented as a fact.

Why the risk axis is often the CFO's argument too

It is worth saying explicitly, because it changes how the room hears the slide: the risk axis is not only the audit committee's concern. It is frequently the stronger financial argument, and a strategist can hand it to the CFO as one. For a company large enough to remain in CSRD scope, the consequence of a restatement, re-assurance, management distraction, and the reputational and financing effects that follow, is measured in a range whose lower bound already dwarfs a year of analyst-hour savings. When you multiply even a modest, defensible probability reduction against a consequence that large, the expected avoided cost can be the single biggest number on the slide. The CFO who came for cost-to-report savings discovers that the more valuable line is the one the audit committee cared about. That convergence is the whole reason the dual-axis frame is honest rather than a compromise: on a well-built program, the two readers are looking at two faces of the same value, and the risk face is often larger.

This also disciplines how you talk about the efficiency axis. If you know the risk axis carries most of the value, you are under less pressure to inflate the speed number, which is exactly the pressure that leads presentations into overclaiming. You can present a modest, tested, honestly-net efficiency figure without anxiety, because it is not carrying the case alone. Paradoxically, the strategist who understands that assurance-risk reduction is the larger prize is the one least tempted to exaggerate the speed gain, and therefore the one the room trusts most on both axes.

Not overclaiming: the sentences that keep you credible

Overclaiming is how reporting AI programs lose the room, and the assurance partner is listening for it specifically. There are three overclaims to refuse on principle, and refusing them out loud is what earns the credibility that carries the decision.

The first overclaim is the false-precision risk number. Never say "this reduces our restatement risk by four percent." Say "on a judgement basis, a cleaner basis of preparation lowers the annual probability of a restatement event by an estimated three to six percentage points, and here are the mechanisms." The range and the label are not weakness; they are the signal that you understand the difference between a measured number and an estimated one, which is the exact distinction the assurance partner cares about most. The second overclaim is the transferred-accountability claim. Never imply that buying a platform moves the obligation to the vendor. Say plainly that accountability for the numbers stays with the company and with management no matter what the vendor's contract says, because the moment you suggest otherwise, the audit committee stops trusting the rest of the slide. The third overclaim is speed sold as if it were free of risk. Never present a cycle-time gain without the risk axis beside it, because a speed gain bought by filling data gaps with unverified estimates is not a win, it is a deferred restatement, and presenting it alone is the single move that most alarms the risk reader.

The cardinal rule sits under all three. Every figure traces to evidence, and "the AI estimated it" is not evidence. When you present to the C-suite, you are not just asking for money; you are demonstrating that the team asking for money understands that rule better than the vendor who sold the tool. The presentation itself is a piece of evidence about whether you can be trusted with the program.

Worked example: the one-view slide

Here is the dual-axis view for an illustrative in-scope company, the single slide you would put in front of the CFO and the audit committee together. Every number is illustrative and must be verified on your own data; the point is the structure and the honesty, not the figures.

Present this slide top to bottom as one story. The efficiency axis says the cycle is materially shorter and the cost is roughly neutral in year one and positive after, with the savings already net of oversight and platform cost, so the CFO sees a return that has survived scrutiny rather than a gross claim. The risk axis says that even before the efficiency case turns positive, the expected avoided cost of a disclosure failure runs to a range whose central estimate can exceed the year-one efficiency position, because the consequence for a company this size is so large. The audit committee sees exposure reduced, priced as a range, tied to mechanisms it can interrogate, with no false decimal to distrust.

Then deliver the recommendation as a conditional, because a conditional recommendation is more credible than an enthusiastic one. Fund the program because both axes move favourably, and only because both do. State the condition explicitly: if the proof of concept had shown the tool closing data gaps with unverified estimates, the risk axis would move the wrong way, the expected avoided cost would become expected added risk, and the correct answer would be to decline despite the speed gain. Hand the audit committee the mechanisms behind the probability reduction and the CFO the subtractions behind the net saving, and let the decision rest on both readings of one view. A strategist who presents the condition, rather than hiding it, is the one the room believes.

Handling the questions you will actually get

Prepare for the three questions that come back every time. The CFO will ask why the efficiency number is not bigger, because a vendor promised more; answer that the number in the slide is the one your own data produced in a proof of concept, and that you would rather present a figure the auditor cannot dispute than one the brochure promised. The audit committee will ask how confident you are in the probability reduction; answer that it is a judgement, presented as a range for that reason, and walk them through the mechanisms that make even the low end defensible, so they are weighing an interrogable estimate rather than a guess. And someone will ask whether the vendor stands behind the numbers; answer that no vendor absorbs the company's accountability, that the obligation stays with management under every framework in scope, and that the program is designed precisely so the company can defend its own figures rather than lean on a supplier's assurances. Each answer is the same move: refuse the flattering version, give the honest one, and let the honesty be the argument.

Key Takeaways

• The C-suite has two readers with opposite instincts: a CFO reads for return (cycle time, cost-to-report) and an audit committee reads for exposure (the probability and consequence of a wrong disclosure). One view must answer both, or one reader hears only half a story.
• A single-axis story fails both. Efficiency alone sounds like accelerating toward a cliff to the risk reader; risk reduction alone sounds like a compliance tax to the CFO. The dual-axis view is the only honest frame, because a genuine win moves both.
• Present the efficiency axis as cycle time and net cost-to-report, with the boundaries frozen, the time-saving tested on your own data not a vendor brochure, and the oversight and platform cost visibly subtracted. A net number that survived scrutiny beats a bigger gross number.
• Price the risk axis as expected value and a range: the consequence of a restatement and re-assurance event multiplied by the estimated reduction in its annual probability, labelled as judgement. For the largest in-scope companies the consequence is so large that a modest probability reduction can rival or exceed the entire efficiency saving.
• Tie the probability reduction to named mechanisms (reconstructable basis of preparation, every figure traced, consistent methods, labelled provenance) so the audit committee can interrogate it rather than take it on faith. An interrogable claim persuades; a bare assertion does not.
• Refuse three overclaims out loud: the false-precision risk decimal, the idea that a platform absorbs the company's accountability, and speed presented without its risk axis. Refusing them is what earns the credibility that carries the decision.
• Every figure traces to evidence and "the AI estimated it" is not evidence. The presentation itself demonstrates whether the team can be trusted with the program, so accountability stays visibly human throughout.
• Deliver a conditional recommendation: fund it because both axes move favourably, and only because both do. A strategist who states the condition, rather than hiding it, is the one the room believes.