It was my second Accountex. The first was exactly ten years ago, in 2016, when every conversation on the floor was about cloud accounting and the looming shift to digital tax filing under MTD. Stepping into ExCeL this year felt like a light-years leap forward. All eyes and ears now were on AI, and watching a dinosaur-costumed robot jump, wag its tail and chase visitors around the hall, I had a question rattling around in my head: do you fear AI, or are you excited by it?
For me, it's definitely the second.
From "just talk" to embedded reality
A couple of years ago — and even as recently as last November — AI in finance was still mostly buzz. Marketing slides, panel chatter, the occasional aspirational demo. Real applications were thin on the ground.
Six months on, the picture has flipped. AI dominated the programme and sessions touching on AI were packed — and these weren't speculative talks. They were live demos, integration walkthroughs, and case studies from practitioners already running with the tools.
The underlying technology has matured fast. The frontier large language models — ChatGPT, Gemini, Claude — are now capable enough that the question is no longer "can this work?" but "where in my workflow does this fit?"
To help people work effectively with those models, a surprising number of sessions focused on prompt writing — the practical craft of telling an AI what you actually want. The gap between someone who can prompt well and someone who can't is now visible in the quality of the output, and the profession is starting to take that seriously. Those who say AI doesn't work apparently need to master exactly that skill.
The next commonly mentioned theme was the use of AI agents. Not just assistants — AI-powered workflow coordinators that can take sequences of actions, make decisions inside defined parameters, and handle multi-step processes end to end. This is the next leap, and early adopters are already building with them.
So: what can humans actually do?
I liked the question displayed provocatively as the introduction to one of the sessions: what can humans actually do that AI can't?
The good news for those who fear is that the answer is more reassuring than threatening.
My question is: do you enjoy running in your never-ending month-end hamster wheel? For years a significant share of finance work has been exactly that — data entry, reconciliations, report formatting, chasing the same numbers in the same templates. Necessary, but not where the judgment lives. And there was no time for actually being a human: standing back, using judgment, applying critical thinking, providing recommendations, daring to challenge and innovate. Do you feel alive only reading that?
AI is taking the wheel. What's left is the part of the job that needed a human in the first place — interpretation, relationships, strategy, the ability to read context and make a call. Human can finally be human in finance and accounting. No more boredom in the role. Everyone becomes a boss with a highly skilled and knowledgeable assistant.
A caveat worth naming
There's a piece of this I think the cheerleaders miss, and it deserves honesty.
Some people will genuinely miss the routine. Repetitive tasks quiet the mind — the way learning a piano piece and playing it every day does, or tending a garden through the seasons, working through the same gym routine, or walking your dog and nodding at the same fellow dog mums and dads each morning. That rhythm has its own value, and pretending otherwise is dishonest about how some of us are wired.
But here's the better trade: instead of running routine reports, take the extra time and walk the dog for a bit longer. Then bring that relaxed mind into the work that actually needs you — the conversations, the calls, the trade-offs AI is not going to make for you. If that doesn't appeal, you might want to retrain as a professional dog walker — because indeed, AI is coming for your routine finance job.
The software race: a sprint to embed AI
The main players in the UK market haven't changed: Sage, Xero, and QuickBooks still dominate, with FreeAgent and Capium among the notable runners.
What has changed is the basis of competition. The race now is to embed AI most effectively into existing workflows. And here's the interesting part: most of these tools are building on the same underlying foundation models, accessed via API from OpenAI, Google, and Anthropic — the same labs behind ChatGPT, Gemini, and Claude. The intelligence itself is essentially a commodity. The differentiation will come from how well it's woven into the specific workflows finance professionals actually use day to day.
QuickBooks captured the right philosophy with their messaging at the event:
"Automation where it counts. Human when it matters."
That's the balance the whole profession is trying to land on.
What else is on the agenda for the year ahead
Beyond the headline AI conversation:
Making Tax Digital (MTD). A decade after it was first announced, MTD is finally live. After ten years of consultation, delays, and staggered rollout, MTD has stopped being a future event and started being a live operational reality. The exhibition floor reflected that. MTD solutions were everywhere — from specialist providers built around the new requirements, to mid-market vendors, to the big names like Sage running dedicated sessions on what compliance now looks like in practice. The breadth itself was the story: the software market has caught up to the regulation, and practices now have plenty of choice.
E-invoicing. The shift to structured digital invoicing is accelerating, pushed by regulatory pressure across Europe and the practical efficiencies it unlocks for firms that move first. UK consultation is ongoing but the direction of travel is clear — next year's Accountex will likely have a much bigger e-invoicing footprint than this one did.
What this means for finance professionals
I left Accountex not anxious, but excited. The profession is in the middle of a transformation that strips away the parts of the job that were always a means to an end — and amplifies the parts that require real human capability.
For the fractional FD model specifically, this is a tailwind. Experienced financial leadership becomes available without the overhead of a full-time hire, and the tools now in a senior finance professional's hands mean the value delivered isn't constrained by capacity the way it once was.
If you're in finance and still on the fence about leaning in — my honest view is that the time to explore is now. Not because the technology is perfect, but because the direction of travel is clear and the upside of moving early is real.
I'd love to hear how others in the profession are approaching this. What's your experience been?
