Blog, Tech: 27.10.26Continuous accounting: what technology changes for an SME owner

For the vast majority of Swiss SME owners, the only clear window on their figures today is still the closing: delivered monthly for the best-equipped, more often once a year for the others. Between two closings, a blur. At Synergix, that horizon moved a long time ago. Accounting runs continuously, thanks to a proprietary platform and automated processes we built ourselves, on a deterministic foundation. We call it continuous accounting. Here is what it changes for a business owner, and why few players in the French-speaking Swiss market operate this model today.

The model still dominant in the fiduciary profession is organised around widely spaced moments of truth. For the best-equipped SMEs, a monthly close that catches up on the previous month’s invoices, reconciles the accounts and delivers a preliminary balance sheet around the 15th or 20th of the following month. For the others, a single annual appointment, the year-end closing, sometimes delivered several months after 31 December. Between two closings, the owner steers by their own dashboard, feedback from their banker and the intuition built up in their trade.

Three structural consequences follow from this model. First, delayed steering, where a decision taken in March relies on a snapshot from January at best, and from December two years earlier at worst. Second, a relationship with the closing experienced as an administrative appointment rather than an operational tool, because it arrives too late to correct anything. Finally, tax planning calibrated after the fact, with adjustments sometimes landing eighteen months after the annual closing.

Technology does not create the demand for steering, it responds to it. What is changing today is what a fiduciary is able to deliver throughout the year, provided it has built the infrastructure behind it. For us, this shift is not a goal to reach, it is the organisation that has been in place for several years.

At Synergix, an accountant’s day does not start with catching up. Incoming invoices are reviewed as they arrive. Account reconciliation is handled during the month rather than piled up until the closing, and the client is contacted when a supporting document is missing, not three weeks later. Month-end is no longer the time to catch up on everything, it is the time to verify.

This daily routine does not depend on plugging in a piece of software. It depends on proprietary infrastructure, built block by block since 2013, and on an architectural approach that sets out what must be automated with deterministic code and what benefits from being handled by a targeted AI model. Three components shape day-to-day work at Synergix.

IODD, the proprietary platform. IODD is the secure workspace shared between our clients and our team. It has existed since 2013 and has never stopped evolving. On the accounting side, it provides access to the file’s data, drag-and-drop invoice upload, connection to Swiss bank accounts and multi-currency management. On the HR and payroll side, it gives access to payslips, salary certificates and absence management. On the treasury side, a centralised view of accounts and a cash flow forecast. IODD is the owner’s interface with their file, and it is what makes continuous accounting possible on the user side.

Management dashboards. Alongside IODD, we provide standardised and customisable dashboards that an owner can use without technical expertise. The tool is separate from IODD but follows the same logic: giving the owner a direct reading of their figures, in step with the month.

A deterministic back-office foundation. What runs in Synergix’s back office is mainly deterministic code: encoded accounting rules, automated controls, approval workflows, coordination between components, archiving and audit trails. This is what guarantees that a calculation reconciles, that a control always produces the same result on the same data, and that a file stands up to any external review. In this infrastructure, AI is not a central block that does everything. It is one component among others, used where it provides what code cannot.

Our supplier invoice processing chain. Where AI makes the difference is on high-volume tasks with highly variable input. At Synergix, our internal processing chain handles around 2,000 supplier invoices a month for several dozen clients. Structured extraction (supplier, amount, IBAN, VAT, reference) relieves our accountants of re-keying and frees up time for checks and advice. Even here, the foundation that holds everything together remains deterministic: coordination between components, validation of outputs, routing of ambiguous cases to an expert, archiving of trails, all governed by code whose logic is explicit and reproducible. AI models propose, the deterministic foundation decides and records. It is one of the AI components of our infrastructure, not the whole of it.

Automated monthly controls. Every month, our clients’ accounts are scanned to flag the entries that need to be justified before the closing. The expert no longer steps in to detect, but to decide on the cases flagged.

Taken together, these components turn the accounting process from an annual event into a continuous flow. Annual accounts remain mandatory under article 958 of the Swiss Code of Obligations, but they become the summary of work already done, not a twelve-month catch-up.

What we have just described is not theoretical. It is operated at Synergix, today, for the files entrusted to us. We mention it because it is also, to our knowledge, a model that few fiduciaries in French-speaking Switzerland operate at the same level of integration.

The reason lies in the order of construction, and in a vision that long predates the arrival of AI in mainstream conversation. Enrico Chincarini founded Synergix in 2001 with the idea of creating a Geneva fiduciary where clients would no longer be kept at a distance from their own figures. Jérôme Chincarini joined him in 2009, convinced that the profession would have to reinvent itself through technology. He created IODD in 2013, at a time when the Swiss fiduciary market was not asking for a platform and building your own was a gamble. Thirteen years later, the question is only just starting to be asked in the profession.

While most traditional fiduciaries scale through hiring (more files, more accountants, and technology coming afterwards in the form of off-the-shelf software), we did the opposite. We started by building our own technology, then ran our finance function through this infrastructure. Software scale and professional responsibility in the same place.

This is not a marketing claim, it is a different production model, built on a vision sustained over time. It explains why an owner who entrusts us with their file is not buying software to operate themselves, nor the hours of an accountant re-keying by hand. They are outsourcing their finance function to infrastructure that is already built and running, and to an expert who puts their signature to the result.

For the owner, the gain is not about accounting, it is about decision-making.

Visibility on the past month: the owner steers on figures from the month just ended, variances against budget are identified without waiting for the annual closing, and a treasury adjustment is made on fresh figures rather than on a projection.

Tax planning ahead of time: decisions that depend on the annual result (dividends, provisions, depreciation, the salary-dividend trade-off) are prepared from October on a solid estimate, not in April on a final closing. See also our article on outsourcing payroll, another service run continuously.

Better-calibrated decisions: launching an investment, hiring, renegotiating a supplier contract or reviewing a pricing policy is done with data from the previous month, not the previous financial year. The owner regains a responsiveness that, until recently, only very large groups with in-house management control enjoyed.

This is precisely the thinking behind our outsourcing services: a back office run continuously rather than a provider called in at closing time.

Continuous accounting is not just a matter of buying modern software. Three conditions separate a marketing promise from an operational result, and they define the architectural approach we apply at Synergix.

Deterministic and probabilistic discipline. At Synergix, deterministic methods govern everything that must reconcile and remain defensible: calculation, control, audit trail. Probabilistic methods come in where judgement creates value: reading, summarising, synthesising, interpreting a complex case. The two coexist and never merge, because merging them is precisely what weakens a fiduciary file.

Control alongside automation. The more a flow is automated, the more it must be controlled automatically. A monthly scan that flags the entries to justify is not a luxury, it is the necessary counterpart of a pipeline that runs without re-keying. Automating without controlling turns a one-off error into a systemic one. We describe this control layer in detail in a dedicated article on monthly closing controls.

The expert at the end of the chain. A Synergix closing is supervised by an expert, tax advice is given by an expert, and a decision that engages the mandate’s responsibility is taken by an expert. The infrastructure executes and alerts, the human decides and commits. This hierarchy is not a concession to tradition, it is what makes the whole defensible in an audit. We have already set out this approach in AI and fiduciary work: where to automate, where to keep the human.

Continuous accounting changes the owner’s position, but not their job. They still have to decide, arbitrate and sign. It changes when they do it, and the quality of the information they rely on. An owner who ran their company on instinct in 2015 can run it on figures in 2026, provided they have alongside them a fiduciary that has built the infrastructure, not just bought the software. Internally, we sum it up in one sentence: accounting follows the business, not the calendar.

It is also the method that Jérôme Chincarini and Loïc Mancino pass on to peers at ESM Geneva, as part of the AI for fiduciaries course (in French) they co-lead with Charles Foucault-Dumas. The next session takes place on 15 and 16 April 2027: two days in person to share what the shift concretely changes in a modern fiduciary file.