Accountex London 2026: Four Practical Lessons for Finance Teams

I went to Accountex London expecting to hear about AI. What I actually took away was simpler — and more useful.

Across every session I attended, one theme kept surfacing: technology only creates value when the underlying finance processes, controls, and operational ownership are already functioning well.

That's not a new idea. But hearing it said, repeatedly, by people who build and sell the technology was striking.

Even the vendors are telling you to fix the process first.

Here are four practical lessons I'm bringing back for the businesses I work with.

1. Your finance "ecosystem" matters more than your accounting software

The first session I attended focused on how finance teams are evolving beyond a single accounting platform — and what that means in practice.

The core argument: most businesses underinvest in the operational layer around their accounts. Purchase processing, invoice approvals, expense capture, credit control — these workflows touch every department, but nobody owns them end to end. Finance ends up as the backstop for everything that slips through the gaps.

The practical implication is straightforward. If your team spends more time chasing missing attachments, late expense claims, and unapproved purchases than they do on actual finance work, the problem isn't your accounting software. It's the absence of a proper upstream process.

What this means for your business: A well-designed purchase-to-pay process — with clear ownership, defined approvals, and consistent controls — reduces the firefighting that consumes finance teams. Month-end becomes a controlled process rather than a scramble. And your bookkeeper or finance manager can spend their time on work that actually adds value.

2. Real-time data is only useful if your bookkeeping is current

The second session made a point that sounds obvious but isn't always acted on: if your bookkeeping is only reconciled monthly, every report you produce is already out of date before anyone reads it.

The tools to enable near-real-time finance have existed for years — bank feeds, automated data capture, coding rules. But many businesses pay for them without building the habits and workflows that make them useful. The data is there. The discipline to keep it current often isn't.

And this matters more now, not less. If you want AI to give you meaningful answers about cash, margin, or supplier spend, the foundation has to be solid. AI applied to stale or poorly structured data gives you confident-sounding answers that are simply wrong.

What this means for your business:Before investing in analytics or AI tools, ask a simpler question: is your bookkeeping actually up to date? Is the data clean and consistently coded? If the answer is no — that's the fix. Everything else builds on top of it.

3. AI is already changing finance work — but foundational AI beats cosmetic AI

The third session drew a distinction I found genuinely useful: the difference between foundational AI (built into how the work actually flows) and cosmetic AI (features bolted on that don't change outcomes).

A lot of what's being marketed as "AI for finance" falls into the second category. It looks impressive in a demo. It doesn't change how invoices get approved, how suppliers get onboarded, or how month-end actually runs.

The session also raised something important about risk: many finance professionals are already spending significant time correcting errors produced by generic AI tools. The output looks authoritative. It isn't always right. And in finance, the cost of a confident wrong answer can be high.

What this means for your business: AI is worth adopting where it genuinely reduces effort and risk — automated data capture, duplicate invoice detection, exception flagging.

But the right question isn't "are we using AI?" It's "is our process designed well enough that AI can help rather than paper over the gaps?"