Accounting software development for the capture, reconciliation and reporting around your ledger.

Buy the ledger. Almost nobody should write their own general ledger, tax engine or statutory filing route: the rules are settled, the packages implement them correctly, and a subtle error surfaces at audit rather than in a bug report. Building starts one layer out, where the work is specific to you: document capture, transaction matching, approval thresholds, and what the monthly pack has to say. The payback is fastest where volume or a late close costs the most hours, so that is where we start.

Yes, and it is the majority of this work. We treat the accounting package as the system of record and build around it: data in through its API or import route, postings back, and the two reconciled so one number never exists twice. The same goes for the CRM, billing, payroll and the ERP. Where a package exposes no usable interface, that layer is the first thing we build, because integration surprises are the ones that move a date.

Three engagement models: time and materials billed hourly and quoted per project, fixed price quoted per project against an agreed outcome, and team extension billed monthly per team member. Which fits depends on how firm the scope is, and we recommend one after discovery. Team extension carries no recruitment or placement fee, and fixed-scope work comes with a post-launch warranty agreed in the contract. We price against your scope, because a bank-feed integration and a full close-and-reporting platform are different projects. Most engagements start within two to four weeks.

Most engagements start within two to four weeks of signature. Duration is governed by your calendar more than by the engineering: a finance team cannot absorb a release during a close, so we sequence around period end and aim to have one workflow running against real data before the next one. Anything that touches the numbers then runs in parallel with the existing process for a full period, and we reconcile the difference before the old path is switched off.

NDAs are signed before anyone starts. We work from least-privilege access agreed with your technical contact, read-only where the work allows, and no write access to production during discovery. Delivery runs with ISTQB-certified QA engineers, and unicrew holds ISO 27001:2022 and ISO 9001:2015, audited by Quay Audit UK. We scope your accounting obligations at discovery: an immutable audit trail behind every posting, e-invoicing mandates, revenue recognition under IFRS 15 or ASC 606, and GDPR over payroll and supplier records.

Yes. It is how our published accounting engagement started. A London financial advisory firm had a partially built MySQL, ASP.NET MVC and C# solution and a partner that had stopped. We picked it up rather than starting again, and added the reconciliation automation, encryption and secure authentication the firm needed. A takeover starts with a read of the code and the data model against what the software is meant to do: the expensive surprises are the assumptions nobody wrote down.

In reading and classifying what already exists, not in generating anything. We have shipped two: document recognition, turning invoices, receipts and statements into postings an operator confirms, and natural language processing over transaction descriptions so they can be categorized instead of eyeballed. Both leave a human in the approval path. Generating narrative on top of a ledger is much weaker, because finance has to say where every number came from. See AI integration services.