Skip to content

Services

AI automation

AI automation for recurring work with documents and enquiries, with an audit trail.

We use AI to automate processes in which people currently read, retype and forward: invoices, enquiries, quotes, incoming mail. Language models handle the reading and suggesting, your staff approve. The result is a business application with logging and permissions, not an experiment.

Typical triggers

Typical triggers

Clients get in touch when one of these situations arises.

  • 01Skilled staff spend hours retyping, sorting and chasing.
  • 02Volume is growing and additional staff cannot be found.
  • 03A first attempt with ChatGPT or a workflow tool showed what might be possible, but is not reliable enough for daily operation.
  • 04Management wants to use AI without letting data out of its hands.

AI automation

What you get

  • Measurement on your data

    Before implementation we test the accuracy rate on real invoices or enquiries. You see figures rather than promises.

  • Workflow with human approval

    Confident cases go straight through, uncertain ones go to the person responsible, with the reason.

  • Integration with ERP, bank and mailbox

    Results end up where you already work, not in yet another tool.

  • Audit trail

    Every extraction and every approval is logged and can be traced later.

  • Data protection concept

    Set out in writing: model, processing region, retention, no use for training.

Typical use cases and effort

Use casePrice rangeDuration
Pre-sorting incoming mailCHF 12,000 to 30,0004 to 8 weeks
Creating quotes from enquiriesCHF 20,000 to 50,0006 to 12 weeks
Invoice capture with bank reconciliationCHF 25,000 to 60,0008 to 14 weeks
Guide values excluding VAT. The price becomes binding after the discovery phase.

What the AI does and what stays in your systems

Language models read documents, classify messages and write drafts. But they do not calculate reliably and do not know your prices. Everything that has to be exact therefore stays in your systems: amounts, terms, stock levels.

Only cases where an error is cheap or certain to be noticed go through without a check. Everything else is approved by someone on your team.

AI automation

Approach

  1. 01

    Intake

    Documents and messages arrive by email, upload or scan.

  2. 02

    Extraction

    A language model reads the content, matches it and rates how confident it is.

  3. 03

    Matching

    Master data, prices and open items come from your systems, not from the model.

  4. 04

    Approval

    A person checks the uncertain cases. You set the threshold.

  5. 05

    Posting

    The result is filed in the ERP, the mailbox or the portal, and logged.

Effort

Price range

Price range
CHF 12'000 to 80'000
Duration
4 to 14 weeks

Non-binding estimate, excl. VAT.

What drives the effort

The key factors are the number of document types, the quality of the incoming data and the depth of integration. For most SMEs, running costs for the models are in the low hundreds of francs per month.

Technologies

  • Azure OpenAI
  • Claude
  • n8n
  • Python
  • TypeScript
  • PostgreSQL

AI automation

Frequently asked questions

How much does AI automation cost for an SME?

A single process such as pre-sorting incoming mail starts at around CHF 12,000. Quotes from enquiries typically cost CHF 20,000 to 50,000, invoice capture with bank reconciliation CHF 25,000 to 60,000. On top of that come model costs, which for most SMEs are in the low hundreds of francs per month, and 15 to 20 per cent of the project costs per year for maintenance and operation.

How do you stop the AI from making mistakes?

By never letting it decide alone where mistakes are expensive. Every extraction is given a confidence score, uncertain cases go to a person for approval, and every decision is logged. Before go-live we measure the accuracy rate on your real data.

How reliably does the AI read our invoices?

Header data such as supplier, amount, date and reference is read very reliably from clean PDF invoices, less so from scans and handwriting. That is why, in the discovery phase, we measure the accuracy rate on a few hundred of your real invoices. Anything below an agreed confidence score goes to your accounts team for checking.

Where does our data go when AI is involved?

We set that out in writing for each project. The default is processing in a European region, under contracts that rule out any use of your data for training. For sensitive data, self-hosted models are an option. Hosting and database are located in the EEA, and processing is designed around the GDPR and the Swiss Federal Act on Data Protection.

Do you work with n8n or build your own software?

Both, depending on the case. n8n suits manageable workflows between existing systems, and your team can help maintain it. As soon as audit trails, permissions, high volumes or a dedicated user interface are needed, a purpose-built application is more robust. We often combine the two.

Does the EU AI Act apply to Swiss companies?

Yes, if you offer AI systems in the EU or the output of your AI is used in the EU, for example a chatbot used by customers in Germany. This is set out in Article 2 of the regulation. If you only use AI internally and for Swiss customers, you are generally not directly affected. Switzerland itself is not adopting the AI Act; it is making targeted changes to its own laws instead.

When do you advise against custom software?

Whenever a standard product covers 80% of the case and the remainder is not a competitive advantage. For accounting, simple appointment booking or a small shop, there are good off-the-shelf solutions. We will tell you so in the intro call. Custom software pays off where several systems have to work together or your process departs from the norm.

Let us talk about your case.

20 minutes are enough to find out whether a custom solution is worth it for you.