The US small-firm tech stack, by stage

Stage one: the sole practitioner, roughly to US$250k

What hurts: the practitioner is the firm, and the firm is a tax season. Client records arrive by email, text and paper, the organizer goes out late, the extension list grows, and the summer is spent on the returns that did not make April. The security plan the tax authority requires exists as a document nobody has read, and client data lives in an inbox.

What works: a tax preparation tool with e-file, a client portal that collects documents and returns the e-signature on the authorization form with the identity check the rules require, a ledger for the bookkeeping clients (usually the one they arrive on), bank feeds and receipt capture so nothing is keyed, an engagement and payment tool so the fee is agreed before the work and collected by card or bank debit, and a written security plan that the stack actually implements: encryption, multi-factor sign-in, backup. Six tools, all priced at a scale one person can carry.

The one thing to do now that pays off later: a standard organizer, a standard engagement letter and a standard client folder, the same for every client. Standardization is free at fifty clients and expensive at five hundred.

Stage two: the first hires, roughly US$250k to US$1m

What hurts: the founder's memory no longer scales, and tax season is now two or three people's crisis instead of one. A second preparer cannot see which returns are waiting on the client, which are in review, which are extended. Bookkeeping clients are served between returns, unevenly. Billing follows the founder's attention and cash follows billing.

What works: this is the stage for a real practice-management or workflow system, because the return status, the deadlines and the client record have to leave the founder's head. US workflow tools are built around the tax season pipeline, and the useful ones show every return's stage at a glance and generate next year's work from this year's. It is also the stage for a document system with a filing convention and retention rules, and for a decided client accounting services offer: what the firm does monthly for a bookkeeping client, at what rhythm, with which tools. The US ecosystem is dense here: several practice and workflow systems, each with a stack of tools that connect to it, and the hub chosen now decides which tools are easy for the next three years.

The mistake at this stage is buying the workflow system for the founder rather than the team. Choose on whether the new hire can take a return from organizer to e-file from it without asking.

Stage three: the firm with a team, roughly US$1m to US$3m

What hurts: capacity. The firm is busy every week and cannot say yes or no to the next client with any confidence. Tax season is survived, not planned, and the October deadline arrives as a smaller copy of April. Client accounting services are the growth line but are delivered by whoever has time, with whatever reporting tool they prefer. Sales tax across states and payroll for multi-state clients turn into research projects.

What works: capacity planning in the practice system, used weekly, with both deadlines modelled rather than absorbed. A reporting and forecasting layer chosen once for the firm, with templates, so an advisory engagement looks the same whoever delivers it. Bookkeeping automation with firm-level rules, so the margin on monthly fees holds. Payroll and sales tax as decided service lines, run through a provider or referred, not improvised. A proposal tool wired to the practice system so accepted work creates the jobs and the billing schedule. And an onboarding sequence that runs the engagement letter, the portal invitation and the security disclosures in one pass.

This is also where AI assistants start to earn their place: meeting notes filed to the client record, drafting from the file, first-pass coding, and the client questions that arrive during tax season answered from the firm's own knowledge. Judged on the hours recovered per client, not on the demo.

Stage four: the multi-partner firm, roughly US$3m to US$10m

What hurts: the firm now has a practice-level view to build and defend: realization by partner, write-downs by service line, which clients are unprofitable, which staff are over capacity in March and idle in June. The tools that served a team of eight creak at twenty-five. The security plan becomes an audit item for clients, insurers and the tax authority. Consolidation and multi-entity clients arrive and the reporting layer has to handle them, and a first offshore or outsourced team needs the same stack with different access.

What works: the practice system's own reporting used as the management report, or a data layer that pulls practice and ledger data into the firm's own dashboards. Security, backup and identity tooling that can be shown to a client, an insurer or an examiner, with access scoped by role and by geography. A consolidation tool for the group clients. A written position on client data in AI tools. And a rule that every new tool must integrate with the practice system as the client record, or it is not bought.

Stage five: US$10m and beyond

At this point the firm is choosing between staying a large practice and becoming something with a different shape: service lines with their own leads, a client accounting services business that runs like a company, an acquisition or a merger, or private equity, which is active in US accounting and arrives with its own stack expectations. The stack question becomes an integration and governance question: one client record, one document system, one identity layer, with the specialist tools hanging off them. That is a programme, not a purchase, and it is beyond what a directory can decide for a firm.

What moves a US firm between stages

Not revenue, exactly. Three things: the first hire who cannot see what the founder sees; the first tax season the firm cannot plan; and the first client, insurer or examiner who asks to see the security plan working. Each is a signal that the current stack has run a stage too long. Firms that read the signal early change one system at a time, in the summer; firms that read it late change three at once in January and pay for it in April.

What US firms actually run at each stage is the question the directory's Power Lists answer from the stacks firms report, once enough have shared them. Until then, the categories linked from this page are the map, the practice systems by country page shows which hubs serve the US, and the cards are the tools.