The accounting firm software stack in 2026: what each layer actually does

Search for accounting firm software and you get lists that put QuickBooks, Xero, Hubdoc, Dext, BILL and Gusto in one ranked table, as if a firm picks one. No firm picks one. Those products sit in five different layers, most firms run four of them at once, and the money is lost in two specific ways: buying two tools that do the same job, or leaving a layer empty and paying for it in staff hours instead.

Layer 1: the ledger

This is the backbone every other tool plugs into, and it is the one decision that is genuinely hard to reverse. Xero runs $25 a month on Early (capped at 20 invoices and 5 bills) through $55 on Growing to $90 on Established, with unlimited users on every tier and the largest app marketplace in the category. QuickBooks Online runs $38 on Simple Start to $275 on Advanced, charges by user seats, and is the stronger choice for US tax and payroll. Sage Accounting sits below both at $19-75 with a large accountant network behind it and a thinner integration story.

The international-versus-US split is the real deciding factor rather than price, and we take it apart properly in Xero vs QuickBooks Online for accounting firms. If you are weighing the Sage ecosystem specifically, that comparison has its own page too.

Layer 2: receipt and invoice capture (the Hubdoc hole)

This layer reads documents so nobody has to type them. It was the layer most Xero firms did not think about, because Hubdoc came bundled; until Xero shut Hubdoc down on 8 May 2026. What replaced it inside Xero, Xero Files, stores documents but does not extract data from them, which means every invoice returns to manual keying unless you buy something.

Dext is where most volume went: best-in-class line-item extraction, deep Xero and QuickBooks integration, and practice plans at $17.70 per client per month on Essentials or $19.20 on Advanced, typically with a ten-client minimum. AutoEntry is the cheaper alternative, Datamolino the one firms choose when per-document pricing fits better than per-client.

Which of them fits depends on document volume and how your firm bills, and we put the two most-compared options head to head in Dext vs AutoEntry.

Firms still working through the migration should start from the best Hubdoc alternatives for accountants, which covers the full replacement set rather than just the two leaders.

Layer 3: accounts payable and bill pay

Capture reads the bill; this layer approves and pays it. BILL is the default in the US: exception-based AP with approval routing, two-way sync to every major ledger, and a free Spend & Expense module with unlimited virtual cards bundled alongside the paid tiers. It costs $45 per user per month on Essentials, $55 on Team and $89 on Corporate, and (the part firms miss when budgeting) per-transaction fees on top, around $0.59 for ACH, $9.99 for an international wire, 2.9% on virtual cards.

There is a hard constraint worth checking before you scope anything: the paying account must be a US business with a US bank account. Non-US firms can be paid as vendors but cannot run their own AP on it. Melio and Nickel are the alternatives where per-transaction economics or that constraint rule BILL out.

The three-way on cost per payment rather than cost per seat is in BILL vs Melio, with Nickel as the free-ACH alternative.

Layer 4: payroll

Gusto is the common pick for US client work, at $49 a month plus $6 per person on Simple, $80 plus $12 on Plus, and $180 plus $22 on Premium, with a Contractor Only plan at $35 plus $6 billed only in months you actually pay someone. Multi-state teams need Plus, which is the upgrade most firms discover late. The accountant-facing part (the Gusto Pro dashboard, revenue share, CPE credits) is the reason it wins firm-level standardisation rather than the payroll engine itself. It is US-only, with no fit for employer operations abroad.

Against the incumbent, the trade is service model versus scale, which we cover in Gusto vs ADP for accounting firms.

Layer 5: practice operations

The layer firms add last and regret not adding first: engagement letters, scoping, and getting paid. Ignition sits here, tying the proposal a client signs to the billing that follows it. It is the only layer on this list that touches revenue rather than cost, which is a strange thing to leave until the end and yet almost everybody does.

How to sequence the build

• Pick the ledger first and standardise on it. Everything downstream integrates with it, and switching later means re-onboarding every client.
• Add capture second if anyone is typing invoice data: this is where the hours are, and post-Hubdoc it is the layer most likely to be silently empty.
• Add AP third, once bill volume justifies approval routing. Below that volume, the ledger plus a bank does the job.
• Payroll follows client demand rather than firm preference, since the client usually has an opinion and a US-only constraint may decide it for you.
• Practice ops last only because it is the one you can bolt on without disturbing the others; not because it matters least.

The stack costs what it costs. What makes it expensive is the per-client and per-user layers scaling faster than the ledger: a firm that priced its stack at ten clients and grew to eighty is usually paying most of its software bill to capture and AP, not to Xero. That is the line to audit, and it is almost never the one firms look at first.