The expense report came back with a line the finance manager had never seen before: a monthly subscription to an AI writing assistant, charged to a sales director's personal card and claimed every month since March.

Nobody had approved it. Nobody had refused it either, which in most mid-market businesses is how a technology decision actually gets made. The CFO's own position had been the careful one. Let the vendors settle, let peers prove the returns, wait and see. It is a respectable stance and it is exactly what a finance chief is there to hold.

But the waiting only happened on paper. Sales had one tool, operations had another, and an analyst in the finance team was pasting supplier contracts into a third, with no security review and no one deciding which of those uses deserved a budget and which deserved a stop.

Wait and see is still a decision. It hands the choice to whoever is keenest, and the keenest person is rarely the one who prices risk for a living. Which uses get money, which get shut down and which sit inside the governed stack are cost-discipline questions, and they are a CFO's questions whether or not the CFO asked to answer them.

The market will stay noisy, and that is a reason to narrow the question rather than leave it open. Choose the small number of uses where the value is real, the cost is contained and governance can start on day one. Say no to the rest with the same care.

The decision has already been made, and nobody signed it.

"We cannot be cutting costs and asking for a new hire in the same meeting."

A non-executive director said that to a CFO at last year's budget round, and the room nodded, because it sounds like discipline.

"Which half of your job was in the position description when you signed it?"

The director's instinct is a reasonable one, and it has a long pedigree. Strategy decks have drawn growth and cost as opposing levers for decades, staff read any cost programme as the first sign of harder decisions to come, and a leadership team that says it will do both can sound as though it will do neither well. One clear agenda often does look more focused than two.

That held while cost work meant cutting.

Cutting is what a business does after growth has stalled: the hiring freeze, the deferred capital, the across-the-board reduction that nobody chose carefully. It is visible, it is painful, and it sends exactly the defensive signal the director was worried about. Optimising is a different kind of work and it is almost silent. It means paying market for what the business buys, governing the renewals that roll over unread, and removing spend that nobody would defend if asked.

A telecommunications contract that renewed itself. An insurance policy accepted at the broker's first quote. A cleaning agreement older than the current site lease. None of it is dramatic, and in many mid-market businesses it is a more reliable source of money for the hire, the new market or the new product than waiting for revenue to arrive.

So take a funding question into the next leadership meeting. Write down what the business would invest in if it had more cash flow, then ask what standing discipline you run for freeing that cash from inside the cost base. If the honest answer is that you have none, the hire is not being turned down on cost. It has been presented without a source of funds, and those are different problems with different fixes.

The CFOs who keep investing through every cycle are the ones who built a standing discipline for freeing cash long before they needed it.

Left alone, the same director will say the same sentence at next year's budget round, and the same hire will be waiting outside the same meeting.