Finance teams are not short of dashboards, automation tools or AI pilots. What they are short of is time and confidence. Time to close, time to forecast, time to interrogate variance properly. Confidence that the numbers in front of them reflect the business as it is, not as it was two weeks ago.

That gap is why the most useful finance conversations right now are not about whether to “use AI”. They are about whether finance can become faster without becoming riskier.

Learn more about Sage Intacct

This was the undercurrent at Sage’s Intacct Breakfast Briefing in Durban, hosted in partnership with TechCentral under the theme “Driving Finance Forward”. The session explored how agentic AI and cloud-native platforms are shifting finance from a back-office reporting function to a forward-looking engine that can guide decisions in real time.

The AI opportunity gap

Erika More, sales director for Sage Africa & Middle East, shared a statistic that will sound familiar to many CFOs: while 86% of CFOs have adopted AI, only 49% have deeply integrated it into core finance functions.

That is not a minor implementation detail. It explains why so many organisations feel they are “doing AI” yet still cannot answer simple board questions quickly: What changed? Why did it change? What is the risk if we wait?

Light-touch AI can make work feel smoother. Deep integration changes outcomes: close cycles, controls, forecasting cadence and the speed at which finance can challenge assumptions before they harden into strategy.

The message for finance leaders is uncomfortable but useful: adoption is easy to report. Integration is harder — and that is why it is valuable.

From efficiency gains to decision advantage

A helpful distinction emerged between an “AI adopter” and an “AI leader”.

Adopters use technology to do the same work faster: reduce manual capture, automate reconciliations, generate reports more quickly. Leaders use it to change what finance is for: earlier warnings, better scenario planning, clearer accountability and faster decision-making with fewer surprises.

This is where agentic AI comes in — not AI that merely produces content, but AI designed to plan and act towards a goal within defined guardrails. In finance terms, it is less about clever answers and more about doing the tedious work reliably so humans can focus on judgment: interpreting results, challenging narratives and guiding trade-offs.

Architecture is strategy

Most finance transformation efforts are constrained by plumbing: how systems update, how data flows, how teams access information and how many workarounds are required to keep the lights on.

In a live demonstration, Michael Britz, senior solutions engineer at Sage, showed how Sage Intacct is positioned as a cloud-native financial management platform. The point was not “cloud” as marketing. It was cloud as an enabler of a different operating model: fewer dependencies, automated updates and a platform designed to keep teams current without bolting on manual processes.

For finance leaders, this is not about convenience. It is about control. Workarounds create hidden risk: fragile reconciliations, spreadsheet-only truth and key-person dependencies that surface at month-end.

Dimensional accounting

A standout concept was dimensional accounting, framed as an alternative to ever-expanding general ledger structures.

Instead of adding segments and complexity to force reporting to work, dimensional tagging allows transactions to carry context — department, location or project. The practical benefit: more granular reporting without turning the chart of accounts into a maze. The strategic benefit is bigger: when data is structured for analysis, finance can respond to new questions without redesigning the ledger each time the business shifts.

Multi-entity complexity

For groups running multiple entities, the pain is predictable: intercompany mapping, eliminations, due-to and due-from positions, and the slow grind of getting to a consolidated view that the CFO can trust.

Britz demonstrated multi-entity management in a single environment, including automated mapping intended to simplify inter-entity processes. The bigger lesson is about capacity. When finance spends less time stitching reality together, it can spend more time advising the business while there is still time to change course.

Most credible AI use cases also least glamorous

The most persuasive AI moments in the session were not flashy. They were practical: accounts payable automation that digitises invoices via a dedicated mailbox and supports line-level matching against purchase orders; an assurance-style agent that flags unusual journal entries in real time by comparing behaviour against historical patterns; and a close workspace designed to track month-end tasks and surface variances.

If agentic AI is going to matter in finance, it will be here: in the repetitive, high-risk work that steals time and invites error.

5 moves to make finance faster without weakening trust

First, identify where decisions wait for data and quantify the cost of delay. Second, fix data structure before adding intelligence — bad inputs produce confident nonsense. Third, automate where repetition and risk intersect: accounts payable, journals and close discipline. Fourth, treat integration as a finance strategy, not an IT side quest — fragmentation is a control problem. Fifth, invest in capability as well as tooling. Adoption fails when teams are not enabled to change how they work.