Breaking silos, recovering an ERP programme, and piloting AI.

The situation

Four functions, four plans, and none of them the same plan.

I came in to Mezzetta as interim CFO with a mandate broader than finance: business integration across finance, operations, IT and strategy, working directly with the CEO and the board.

The pattern was familiar in shape and unusually costly at this scale. Departments were optimizing locally. Finance, sales and marketing each built their own view of the year, each defensible on its own terms, and none of them reconcilable with the others. The annual budget consumed four months, most of it spent negotiating between versions of reality rather than deciding anything. Projections that emerge from that process are not credible, and everyone in the room knows it.

Two system initiatives compounded it. A major ERP implementation had gone off track. And a trade promotion management decision — the system that governs how promotional spend is planned, executed and measured, which in CPG is close to the centre of the P&L — was being approached without a structured evaluation.

What I did — planning

One plan, built once, by everyone who has to live with it.

Rather than arbitrate between three budgets, I rebuilt the process so only one got made. Finance, sales and marketing now build a single plan together, against the company's strategic goals rather than against each other's assumptions.

• Streamlined the underlying work processes so that departmental activity ties back to shared strategic goals, and the handoffs between functions stopped being places where information was lost or reshaped.
• Set controls at the level the business could actually carry — enough to make the numbers defensible, not so much that the operating teams slowed down to feed the process.
• Cut the budget cycle from four months to one, with the teams aligned at the end of it rather than exhausted by it.

What I did — systems and vendors

Choose vendors by evaluation, not by momentum.

Large system decisions tend to get made by whoever built the relationship first. I replaced that with a structured process, run collaboratively across the functions that would have to use the result.

• Ran a proper sourcing and evaluation process for a new trade promotion management system, bringing the affected departments into the assessment rather than presenting them with a decision. In CPG, TPM is not an IT purchase — it determines how well the business can see the return on its single largest controllable spend.
• Put the stalled ERP implementation (IFS) back on track. Recovering a programme like this is mostly not technical work. It required tough internal conversations about how it had gone wrong: scope that had been defined too loosely at the outset, technical assumptions that had been carried forward without ever being validated, and a budget that had never been sized for the scope people expected to receive.

Those three failures compound quietly. Loose scope invites optimistic assumptions; optimistic assumptions justify a budget that cannot deliver; and the gap only becomes visible once enough money has been spent that nobody wants to be the person who names it. Naming it is the job.

What I did — AI pilot

An internal pilot, run with IT rather than around it.

I designed and implemented an internal AI pilot in partnership with the IT function, aimed at two outcomes rather than one: raising day-to-day productivity, and raising the quality and depth of the analysis the organization is able to produce.

The second matters more. Automation that only makes existing work faster returns hours. Tooling that lets an analyst ask a harder question than they could previously answer changes what the business is able to decide. Running it in partnership with IT — rather than as a finance side-project — is what makes it survivable past the pilot.

What changed

A plan the business believes, and programmes that move again.

The budget process now runs in a month instead of four, and produces projections that finance, sales and marketing all stand behind — which is a different thing from a budget that merely got approved.

The ERP programme is back on plan, with its scope, assumptions and budget reconciled to each other rather than to wishful thinking. The TPM evaluation is being decided on evidence and cross-functional input. The AI pilot is live with IT.

Engagement ongoing — figures to be finalized at close