AI in accounting for the reconciliations, close prep and client paperwork your team still does by hand

Your staff tie out bank statements, chase clients for documents and build the close file line by line. We build AI agents that take the first pass inside your ledger, practice management system and document portal, and an accountant signs off before anything is posted, filed or sent. Client data stays in your own cloud.

Book a Discovery Call Start with an AI Audit

Processing supplier invoices? Capture, matching and posting have their own page: AI invoice automation. Jason, our COO, runs the call and scopes the first workflow with you.

Agent run: September bank reconciliationSample

Statement pulled1,214 lines from the operating account bank feed
Done
Matched to the ledger1,187 lines matched on amount, date and reference
Done
Exceptions grouped27 open: 19 timing items, 6 missing receipts, 2 unknown payees
Done
Client request draftedSix receipts listed by date and amount, ready to send through your portal
Done
Reviewer sign-offProposed adjusting entries and the two unknown payees wait for a senior
Waiting
Filed to the close binderReconciliation, tie-out to the trial balance and run log entry
After approval

What the agent can seeOnly the fields this reconciliation needs. The rest of the client file stays in your systems.

Bank statement lines: Visible
Cash account in the ledger: Visible
Receipts on file: Visible
Client contact: D. H••••
Full bank account number: Masked
Other clients' files: Not retrieved

Illustrative run. Matching rules, thresholds and routing are set per workflow during scoping. The agent never posts a journal entry on its own.

Clients include

AI use cases in accounting, from client intake to month-end close

Start where the volume is high, the procedure is written down, and a reviewer already checks the work before it counts. In a CPA practice or an in-house finance team, these six workflows meet all three. Accounting software already covers some of this, and our build vs buy guide helps you decide which workflows need a custom agent.

Intake

Client document collection

The agent compares what each client has sent with the request list for the engagement, sends reminders for what is missing, and files every upload to the right client and period. Your staff open a complete folder instead of an inbox.

How agents connect to your ledger and portal

Reconciliation

Bank and balance sheet reconciliations

The agent matches statement lines to the ledger, groups what is left by cause and proposes the adjusting entries with the evidence attached. A senior approves before anything is posted.

How approval gates work

Month-end close preparation

The agent works your close checklist: it pulls the trial balance, flags accounts that moved past your threshold and drafts the variance commentary from figures the database computed. Your controller edits and approves the commentary.

How the agent reads your ledger figures

Tax

Tax workpaper preparation

The agent reads the source documents, fills your workpaper templates, ties each figure back to the trial balance and lists what the client still owes. Your preparer and reviewer sign, and the agent never files a return.

Prove it on last season's files first

Onboarding

Engagement letters and client onboarding

The agent drafts the engagement letter from your approved template and the scope agreed on the call, then sets up the client in your practice management system. A partner reviews and signs every letter.

How your firm's rules reach the agent

Client queries

Answers to routine client questions

Questions about deadlines, document status and what to send next get an answer from the client's own file and your procedures. Anything that needs professional judgment goes to the accountant on the account, with the thread attached.

How a second agent checks every reply

The capacity squeeze behind every close

18%

of accountants make financial errors at least daily, in a survey Gartner tied closely to low capacity

Source: Gartner, survey of 497 people in the controllership function, conducted July 2023, published 21 February 2024.

Gartner and the US Bureau of Labor Statistics measured the pressure on accounting teams. Each figure maps to the work an agent takes on.

59% make several errors a month

A third make a few every week. The reconciliation agent ties every line back to the ledger before a reviewer opens the file, so the review starts from the exceptions.

73% say new regulations added to their workload

And 82% say economic volatility increased demands for their work over the past three years. Workpaper and close agents take the first pass on the volume, and your people keep the judgment.

115,300 openings for accountants and auditors each year

The BLS projects that many openings a year through 2035, while total employment grows by 79,400 over the whole decade, so most openings replace people who leave. Hiring alone will not cover a heavier close.

Start with work a reviewer already checks

An agent that prepares a reconciliation or a workpaper is checked by the senior who reviews that work today. The AI Audit ranks your candidates by value, effort and risk, and the AI readiness assessment checks whether your data and systems are ready.

Sources: Gartner, "Gartner Survey Shows That a Third of Accountants Make Several Financial Errors Per Week due to Capacity Constraints" (press release, 21 February 2024); US Bureau of Labor Statistics, Occupational Outlook Handbook, Accountants and Auditors (projections 2025 to 2035).

Where the agent works in your month-end close, and where an accountant signs

Five steps from open books to a reporting pack. The agent takes the first pass at each, and every entry, adjustment and figure that leaves the team carries a person's sign-off.

Collect
AgentPulls bank feeds, statements and open documents, and chases what is missing
PersonStaff accountant confirms the period is complete
Reconcile
AgentMatches every line to the ledger and groups exceptions by cause
PersonSenior approves the proposed adjusting entries
Accrue
AgentDrafts recurring accruals and prepaids from your schedules, with the support attached
PersonAccountant reviews and posts each entry
Review variances
AgentFlags accounts past your threshold and drafts commentary from computed figures
PersonController or manager edits and approves the commentary
Report
AgentAssembles the pack and the close binder, with the run log for each step
PersonController or partner signs off the period

Illustrative. Which steps an agent takes, and which thresholds send work to a person, are set per workflow during scoping.

How an accounting agent handles client data and ledger access

Your partners, controllers and auditors will ask what the agent saw, what it changed and who approved it. We build those answers into the system before it goes live.

Your environment

SourcesLedger, practice management, document portalRead through the interfaces they already expose
Context layerScopes and masksOne client per task, account and tax numbers masked
AgentSelf-hosted n8nRuns the workflow and your rules in your cloud
ApprovalReviewer sign-offEvery entry and client message, before it is posted or sent

run 8812 · bank reconciliation · 3 sources read · account number masked · 2 adjusting entries proposed · approved by senior · posted to ledger

Outside the boundary

A model provider receives scoped, masked fields only, under your own agreement with that provider, or is replaced by a model you host.

Never allowed

An agent posting entries, filing a return or signing on its own
Client data used to train models
Figures the model calculated itself

Sign-off

A named reviewer on every entry

Adjusting entries, workpapers, engagement letters and client messages wait for a named person before anything is posted or sent. Each workflow has an owner and an escalation path before it goes live.

See AI governance

One client at a time

Each task sees one client's fields

A context layer between your systems and the agent limits each task to the client and the fields it needs, and masks bank account and tax identification numbers it does not need.

See the context layer

Your cloud

Every run logged in your own cloud

We deploy on self-hosted n8n in your cloud, behind your own auth, with permissions inherited from the systems the agent reads. Every run records what it read, what it changed and who approved it.

See n8n for enterprise teams

From the first call to an accounting agent in production

One team carries your workflow from scope to the system running in production, and every build is tested in staging before it goes live. See how we work

Jason, COO
AI Audit, if you need one
Eight phases across strategy, operations, data and systems over 4 to 8 weeks, ending in a ranked 90-day roadmap. The fee is credited toward your first build.
Jason, COO
Scope
One workflow, the systems it touches and what it will take, agreed before you commit. Owen joins the first calls until you are onboarded.
- The clients, accounts and periods in scope
- The thresholds that send work to a reviewer
- The timeline, before you sign
Maciej, CTO
Build
Built against your real data, in your environment, as production software from the first week.
- Staging first, on a closed period or test client
- Your staff run it on a live close before go-live
- Self-hosted n8n in your cloud, built by our n8n agency team
Jason, COO
Integrate
Into production behind your auth, your governance and your monitoring, with your team trained on it.
- Separate development, pre-production and production environments
- Runs behind your own auth
- Your people trained on the system
Maciej, CTO
Run and improve
Watched, fixed and extended in weekly sprints by the engineers who built it.
- Error alerts to the engineers who built it
- Weekly sprints that you direct
- The next workflow, scoped with you

What your firm or finance team gets from a build with us

We build the agent around your procedures and run it in your environment. To train your staff on working with it, see our corporate AI training programs. Planning AI across the whole firm rather than one workflow? Start with AI consulting, or read how we approach enterprise AI strategy.

	What you get
You receive	Working agents inside your ledger, practice management system and document portal
Where client data lives	In your cloud, on self-hosted infrastructure
Fit to your procedures	Built around your close checklist, workpaper templates and review steps
Sign-off	By your accountants, approved and logged
Timeline	Scoped per workflow and agreed before you sign
After launch	The engineers who built it monitor and extend it

What accounting firms and finance teams ask about AI agents

Ask us something else

How is AI used in accounting?

Mostly in the preparation work behind each sign-off: collecting client documents, matching bank lines to the ledger, preparing the close, filling tax workpapers and drafting engagement letters. In each one, an agent reads the documents and systems a staff accountant would, assembles the file and drafts the next step, and an accountant makes the call. Fully automated steps are limited to work with written rules, such as reminding a client about a missing statement. Our 25 enterprise AI use cases by department covers the finance and procurement cases beside these.

Will AI replace accountants?

Our builds keep an accountant on every decision. The agent takes the first pass, such as matching 1,187 of 1,214 statement lines in the sample run above and grouping the other 27, so your staff spend their time on exceptions and judgment. The BLS still projects employment of accountants and auditors to grow 5% from 2025 to 2035. Which tasks move to an agent is agreed in the scope.

Can AI do bank reconciliations?

It can do the matching and the first pass on exceptions. The agent matches statement lines to the ledger on amount, date and reference, groups what is left by cause, drafts requests for missing support and proposes the adjusting entries. A reviewer approves before anything is posted, and the reconciliation is filed with its run log.

Can an AI agent prepare tax workpapers?

It can prepare the first draft. The agent reads the source documents the client sent, fills your workpaper templates, ties each figure to the trial balance and lists what is still missing. Your preparer reviews and completes the workpapers, your reviewer signs, and the agent never files a return.

Which accounting systems can an agent work with?

Whatever your systems expose: a vendor API, a database view you give us read access to, scheduled exports, or the documents that arrive by email and portal upload. Cloud ledgers such as QuickBooks Online and Xero publish APIs, and for an older on-premise system, exports into a staging database often work best. Our guide to how agents connect to your ERP and core data covers the patterns, and the AI Audit confirms what each of your systems allows before anything is quoted.

Do you also automate accounts payable invoices?

Yes, and that work has its own page. Invoice capture, extraction, matching to purchase orders, exception routing and posting to your ERP are covered in AI invoice automation. This page covers the work around it: reconciliations, close preparation, tax workpapers and client paperwork.

Where should an accounting firm start with AI?

With the workflow that has the most volume and the clearest written procedure, usually bank reconciliations or client document collection. If you already know which, book a discovery call and Jason scopes it with you. If you have several candidates, the AI Audit ranks them by value, effort and risk over 4 to 8 weeks and ends in a ranked 90-day roadmap. It costs $10,000 to $100,000 depending on scope, and the fee is credited toward your first build. To prove one workflow first, an AI pilot runs 2 to 6 weeks on your own data for $35,000 to $75,000, also credited toward the build.

Who looks after the agents once they are live?

The engineers who built them, in weekly sprints that you direct, with error alerts going straight to them. Busy season changes volumes and rules, so the next adjustment is scoped with you. If another team built workflows you already run, we take them over after a technical assessment of their quality. See AI managed services.

AI for related professional and financial services

Firms that serve clients on engagements, and teams that live in the ledger, share most of these workflows.

Legal

AI for lawyers

Conflicts checks, contract review against your playbook and pre-bills, with a lawyer approving every output.

See AI for lawyers

Consulting

AI for consulting firms

Proposals, research syntheses and decks drafted from past engagements, with a partner approving everything a client sees.

See AI for consulting firms

Banking

AI in banking

KYC checks, AML alert triage and loan file preparation, with an analyst approving every decision.

See AI in banking

Private equity

AI in private equity

Deal screening, due diligence and portfolio monitoring, including the KPI pack each portfolio company sends.

See AI in private equity

Put your first workflow on the board

Answer a few questions about your team and what you want automated, then Jason scopes it with you on the call.

Owen, Maciej, Jason, Oskar and Ben