For years, finance was the organisation’s scorekeeper: close the month, balance the books, keep the auditors happy. That work still matters, but it no longer defines the job. In an economy where costs jump, supply chains wobble and assumptions expire quickly, the modern finance function is being judged on one thing above all: how fast it can turn uncertainty into a decision the business can defend.

That tension sat at the heart of a recent TechCentral and Sage Breakfast Briefing at The One & Only, where CFOs and senior finance leaders compared notes on what it takes to build a future-fit finance team. One phrase landed because it captured the shift neatly: the CFO as “Chief Future Officer”. Not a motivational poster, a practical expectation. Finance is increasingly expected to set direction, not just report outcomes.

The bottleneck isn’t ambition, it’s the forecast

When leaders talk about faster decisions, the conversation often drifts towards technology first. At the briefing, the most persistent blocker was more basic: forecasting and budgeting cycles that are too slow and too rigid for today’s volatility.

If your forecast takes weeks to refresh, it becomes a historical artefact the moment it lands. And when the budgeting process is built around fixed assumptions, the business starts working around finance rather than with it. The result is predictable: leaders either delay decisions while they wait for a number they trust, or they move ahead without finance and deal with the consequences later.

Speed, in other words, is not just a performance improvement. It is a governance issue.

Single source of truth

The second friction point was fragmentation: systems that do not speak to each other, creating silos, reconciliations and manual workarounds. Finance then spends its time stitching reality together, instead of interrogating it.

The phrase “single source of truth” gets used casually in boardrooms, but finance leaders know it is hard for a reason. It requires decisions about ownership, permissions, where data lives and how integrations scale as needs evolve. It also forces uncomfortable trade-offs: do you build a unified view by standardising processes, or do you accommodate complexity and accept slower closes?

One useful reminder from the briefing was that scale is relative. South African organisations range from a handful of finance users on a platform to hundreds. The point is not whether a tool is labelled “SME” or “enterprise”, but whether it can be scoped, governed and extended without collapsing into spreadsheet glue.

AI is only as smart as your messy middle

No modern finance discussion escapes AI. But the tone here was refreshingly pragmatic: the world may be crazy about AI, yet the real question is how it makes finance better on a Tuesday morning.

AI becomes valuable when it shortens cycles, improves forecast quality, spots exceptions earlier or reduces risk in routine decisions. What it cannot do is compensate for poor data quality, weak controls or a process landscape that no one fully understands. If your team cannot agree on which numbers are correct, automating the disagreement just produces faster confusion.

The unglamorous work still wins: clean data, defined processes, clear accountability and a cadence for reviewing performance.

The talent shift: from transaction work to judgment

The capabilities finance leaders are building now centre on upskilling and reskilling. The goal is straightforward: move teams away from transactional work and towards analysis, judgment and strategic support.

That demands learning agility, not only training. It also demands a realistic plan for adoption, especially in organisations where legacy habits, and sometimes legacy systems, are deeply embedded. Finance can design the perfect process and still fail if the business does not use it.

5 practical moves to speed up decisions without losing control

Measure where decisions wait for data and put a cost on the delay.
Define what “truth” means in your organisation: ownership, access, audit trail and accountability.
Pick one manual, high-effort process and automate it end-to-end, then redeploy the time saved into analysis.
Pilot AI against a specific outcome and track whether it improves speed, accuracy or risk.
Treat adoption as part of the deliverable: training, incentives and leadership attention, not a footnote.