DAS 2026 Wrap: General Ledger and the Battle for World Domination

By Damon Anderson, 9 October 2026. 12 minute read.

About six months ago I started the A to Z. I set out to find the most interesting companies in the space, and I was pretty gutted, because none of them were British. Least of all in the general ledger, the master record of every debit and credit at the heart of any accounting system. The ledgers that mattered were almost always foreign-born.

Six months on, the Risk board looks very different. Briefcase and Finzu are building ledgers, Combinely the agents that work across them, and Koovo the advice that comes out of them. All UK players, each in a different part of the workflow, with tiny armies and big ambitions. (Disclosure: I advise Finzu as part of my consultancy business.) And this week Craig Walker of Digits, who wrote the first line of Xero's code in a Wellington apartment twenty years ago, was on stage in London.

It reminded me of a game of Risk.

Risk is the board game where you mass armies on a map of the world and try to take continents. You rarely win by brilliance. You win by attrition, and by keeping your rival busy on someone else's border.

In English-speaking accounting software, the board has two big prizes: the US and the UK. Australia, New Zealand and Canada matter, and great products have come out of all three. But take America and Britain and you've won the game.

Rewind a decade or so. Xero had taken Australia and New Zealand and went after the big two, landing early in the US. That woke Intuit up. As the story goes, Intuit shored up QuickBooks Online at home, then sent troops into the UK, where Xero was already scrapping with Sage. The point was to keep Xero fighting on a second front while Intuit fortified America. Xero went on to lead the UK. The US is still a much harder fight.

Now look at the board. In the US, Digits played the Trojan horse. It lived inside Intuit's ecosystem as a helpful app, earned accountants' trust, then built a ledger of its own. Briefcase did the same here: plugged into Xero, then launched its own ledger with a one-click way out. (More on how long that lasted below.)

The war chests differ too. The venture capital sits in Silicon Valley, and Digits is several funding rounds in. In the UK, most of the money going into accounting software is private equity, buying up existing products and bundling them into roll-ups. Fewer bets on new ledgers, more on consolidating old ones.

So the real question is who wins the UK, where Xero now has loads of troops on the pitch. And one of the people who built Xero is standing on a London stage for Digits. Reconnaissance mission? I'd put money on it. If the armies follow, the UK becomes a proper front: Xero and Sage defending home turf, homegrown upstarts building ledgers, and an American AI-native landing on the beaches.

That was the backdrop to two days at the Digital Accountancy Show.

Two days at the sharp end

DAS is the UK's flagship tech-first accounting event, and next year it hits the US, in Austin, Texas. Yes siree.

It opened with lasers, a moody video, a smoky room and graffiti on the walls. Full Xerocon circa 2019 vibes: punk and street meet accounting, inside the conference factory that is ExCeL. The floor was packed with around 190 exhibitors, silent-disco stages (you listen through headphones) and a programme with some proper gems. My pick was the AI doomsday clock panel with Indi Tatla, Dave Sellick, Stuart McLeod and Becky Shields. The profession arguing with itself in public. More of that, please.

AI is everywhere, and still impenetrable for most

AI was in the title of 49 of the 137 sessions on the DAS 2026 agenda. More than a third.

Two letters on every stage and every stand, and they make it harder, not easier, to tell one product from another. We need more coherence in how we talk about AI.

The test I keep coming back to is AI-native versus AI-added. AI-native means the product was built from day one for AI agents, software that takes actions rather than just answering questions, to do much of the work. AI-added means a product built for humans clicking buttons, with AI bolted on later. Both can be good. "AI-powered" on a banner won't tell you which one you're looking at.

Back to the board: the ledger war

I recorded a podcast on AI-native ledgers with Tom Herbert for AccountingWEB's No Accounting for Tech, live from the show floor. A year ago this was a US story. This week Briefcase was demoing a UK-built AI-native ledger in the same hall Digits was speaking in. The A to Z now tracks about 25 AI-native ledgers worldwide.

What does Xero make of the Trojan horse? We may already have the answer. Since this piece first went out, Briefcase has removed its one-click Xero migration. Search the Xero App Store for Briefcase today and you get nothing back, and its old listing page returns an error. Whether that was Briefcase's call or a nudge from Xero's partner terms, the switch has gone. And there's the irony: Xero's open API, the thing that built its huge ecosystem, is also what makes it easy to lift your data out and leave.

Here's a sentence I never thought I'd write: in the LLM era (large language models, the tech behind ChatGPT and Claude), Xero is the incumbent. It was built for a human with buttons and drop-downs, not for agents. Still, swapping the system at the heart of a practice isn't for the faint-hearted. The new thing has to be simple to switch to or ten times better. Ideally both.

My view: the general ledger sinks to infrastructure, and that's fine. It becomes a deterministic record, one that gives the same right answer every time, that the user may never open.

Win the ledger, lose the war?

Here's the twist in this game of Risk. The ledger is the territory everyone is fighting over, but territory is just infrastructure. The bigger prize is whoever commands the armies: the agent that actually does the work.

Every ledger now faces a choice. Open up through MCP (more on that below) and other people's agents can come in and do the work, which makes you the ground someone else's army marches on. Stay closed and you risk being left out of where the work is heading.

So whose agent will it be? The ledger's own built-in assistant. Claude or ChatGPT, sitting on the accountant's desktop. An orchestration player running work across every system the firm uses. Or the practice's own agents, built around how that firm works. Increasingly it won't be a human doing the clicking, even if today it mostly still is.

It's the question I closed my talk with at the Chift dinner on Wednesday. In two years, a business owner says to an AI: close my month and tell me what I owe. Whose name is on the screen when it's done? The ledger, the app that did the work, or the AI they asked? Nobody has won that one yet, and it's the most interesting fight on the board.

Everyone is building for the same surface

More work now starts in Claude or ChatGPT than in the accounting app. I made this point in the same Chift talk, and the A to Z data backs it up.

The connection that matters is MCP, the Model Context Protocol: a universal adapter that lets an AI assistant plug into software, read its data and, if allowed, take actions. About a quarter of the nearly 400 vendors on the A to Z have one. Only around 30 go deep. Most can see your data. Very few can safely change your books. Fewer still can show you an audit log of what changed.

So everyone is fighting for the same surface, the place the user actually talks to. Today that's still mostly the ledger. Tomorrow it might be Claude. The ecosystem doesn't die. It follows the user.

That raises the question nobody wants to answer on a stand: trust. When a client's books travel from the ledger into a chat window, who can read them on the way? Where is the data encrypted, and who holds the keys? A business has a duty to stop staff going wild on the corporate card, and the same duty for an agent. The vendors who answer that plainly will win the accountants still sitting on their hands.

The plumbing matters more than the demos

To chase an invoice properly, an AI needs to reach the books, the bank and, for real accuracy, the till. An AI is only as good as the systems it can reach. That's why I spent time at the unglamorous end of the hall. Unified API providers build one connection to hundreds of finance apps so nobody else has to, which means they decide what an agent can touch. Chift is the Codat for Europe. If you're building an AI product, build the plumbing or buy it. You can't skip it.

The hard graft is inside the firm

Which brings me back to who commands the armies. Orchestration is where a lot of the real work now happens: engines that sit over a firm's existing tools and run the work across them. Dytto out of Belgium, Ravical out of the UK and Belgium, Artifact and Combinely out of the UK, and the big daddy, Basis, in the US. Combinely runs an AI coworker of more than 50 agents that logs into the ledgers a firm already uses, like a new member of staff.

The best of them have one thing in common, and it's people. Smart people who go into a practice, learn how it really works and where it gets stuck, then build the tech off the back of that. It doesn't fit on a stand banner, but it's where most providers need to point their effort.

Time for advisory to shine?

Craig made a point of not saying advisory is the answer, and I was glad he did. We all said it in the Xero days. Automate the books and all those capable people are freed up to advise. I helped tell that story, but we never fixed the first bit. Now, with a hall full of AI bookkeeping apps vying for your business, it might finally happen.

What strikes me is how little tech innovation there is in advisory itself. Cash flow forecasting apps are about as good as it's got, and a forecast is only as good as the quality and freshness of its data. I've yet to see anyone automatically pull in what's happening in the world outside the ledger.

Then I met Ollie Jackson, founder of Koovo, a recent entrant on the A to Z. Koovo builds an ontology of the physical world, a map of how businesses, products, suppliers and markets connect, from what's in the general ledger. Then it reads the market and predicts where things are heading. Picture a supply chain shock. An adviser who can show a manufacturer when it becomes material to their business, and have that conversation before it bites, is offering something new. That's a new category of work, and of revenue.

That's what makes me optimistic. Forget the doomers, the trolls and the hype merchants. The machines are starting to do work that serves the client better than we could. This was never about protecting the bits of the job we happen to enjoy.

So where should you start booking demos?

Well, not necessarily in the exhibitor hall, at least not on its own. Around 190 stands, and at every second one a bowl of jelly beans and a lovely person whose job is to stop you walking past. In 2026 that's still how our industry expects buyers to find software. That's no slight on DAS. Every conference works this way. I often say searching for software online shouldn't feel like wandering round an exhibitor hall. Standing in one, I'm not sure the exhibitor hall should feel like that either.

Buyers: do your research before you arrive, and make every vendor on your shortlist show you the thing working on your own data. Vendors: be findable on evidence, not stand size. Events are where you meet the people behind the shortlist. They're a poor place to build it.

The battlefields to come: FAB UK and DAS US

If the US has the money, the UK has the events. Always has. Xerocon London, Accountex, DAS and now FAB, the Finance, Accounting and Bookkeeping Show, make Britain the place this industry comes to show its hand.

So circle the next two dates in the diary, because I have: DAS landing in Austin, and FAB at the NEC in Birmingham next March. Both arrive at a critical moment in this game of Risk. I'll be watching who turns up, and how many troops they bring.

Full disclosure: I'm helping the FAB team with their AI programme. It has no bearing on anyone's ranking on the A to Z. More on that here.

Start your shortlist today

The game of Risk is back on, and this time the armies are AI. I went to DAS looking for the next Xero and didn't find it. Nobody has won yet. But I saw where it'll come from: teams doing the hard graft inside firms, the plumbing nobody claps for, and the few brave enough to rethink what advice looks like.

The next Xero won't hand you a jelly bean. It'll show you your own numbers, and something you didn't know about them.

Want a head start? The A to Z is free for accountants and finance teams: the directory, the World Ranking, and a recommendation engine that starts from your problems, not a vendor's pitch.