GKS Solutions

AI & automation for SMEs · Canton of Zurich

Less admin. More business.

We find the repetitive office work that slows your team down every week and automate it with AI: inside the systems you already use, with your approval where it matters, and with results you can measure.

  • Directly with the three founders
  • Your existing systems
  • One process at a time

Example: one supplier invoice

Show process
E-mail invoice arrives invoice arrives Excel list kept by hand updated automatically Accounting typed in again entry prepared Filing PDF filed by hand filed automatically AI reads & captures amount, date, supplier Your approval one click, document alongside open retype look up enter file update Your week Admin Time for the business

Today, you are the interface: every red dot is a person carrying data from one system to the next.With GKS, the work flows between your systems. A person only checks where it matters.

  • A person steps in
  • Automated flow

Simplified illustration, no client data.

01 — The situation

Your business grows. So does the paperwork.

Most owner-led SMEs have no team for IT or process improvement. Those topics land on the management, on top of everything else. And the hours disappear into tasks like these:

  1. A

    Retyping and transferring

    Invoices, delivery notes and forms are copied by hand into accounting, the ERP or Excel, often more than once.

  2. B

    The same questions, again

    Quote requests, appointment requests and status questions arrive by e-mail, phone and WhatsApp and tie up your best people.

  3. C

    Knowledge in heads and folders

    Where is the current price list, which contract clause applies? New staff have to ask and hold others up.

Nobody needs AI for its own sake. What counts is time back, fewer errors and more capacity for your customers.

02 — What we build

Concrete solutions for work that repeats.

We connect AI to the software you already use. Three typical starting points:

  • 01

    Document automation

    Invoices, delivery notes and forms land in the right system without retyping.

    1. Document arrives
    2. AI reads & classifies
    3. Your team checks uncertain cases
    4. Data in the right system
  • 02

    Customer assistant

    Answers recurring customer questions around the clock and hands over when it matters.

    1. Customer asks a question
    2. Assistant answers from your information
    3. Handover to your team when needed
    4. Customer has an answer
  • 03

    Knowledge assistant

    Your team asks in plain language and gets answers from your own documents, with sources.

    1. Question in plain language
    2. AI searches your documents
    3. Answer with source, your team checks
    4. The right information, faster

Your process isn’t listed? We build what it needs.

See all solutions

03 — Do the maths

How much time is in a single task?

A simple scenario with your own assumptions. Not a forecast and not a promise, but a good starting point for the discovery call.

Your assumptions

40 per week

For example incoming invoices or customer enquiries.

8 minutes

Including searching, retyping and following up.

50 percent

Estimate conservatively: exceptions and control stay with you.

CHF 85

Salary plus overheads, roughly.

45 weeks

Excluding holidays.

Illustration Your scenario

Hours per year
120 h
Hours per week
2.7 h
Value per year
CHF 10'200

How we calculate

40 × 8 min × 50 % × 45 weeks ÷ 60 = 120 h

120 h × CHF 85 ≈ CHF 10'200

A scenario based on your inputs, not a commitment. Implementation and running costs are not deducted. We measure real values in the pilot against your baseline.

Discuss your scenario on a call

04 — Approach

One process at a time. You decide after every step.

Five stages from the first question to a solution in daily use. Scope and price are clear before each stage. Afterwards, you decide whether to continue.

  1. Discovery call

    Free · 60 min

    We find the process where automation is most likely to pay off.

    Your decision: Is a closer look worthwhile?

  2. Process check

    Defined scope

    Current process, baseline and expected benefit, in writing.

    Your decision: Do we start a pilot?

  3. Pilot

    Defined scope

    Automated on your real data, measured against today.

    Your decision: Does the solution go into production?

  4. Production

    Defined scope

    Connected to your systems, team trained, properly handed over.

    Your decision: Who runs the solution?

  5. Run & Care

    Optional · monthly

    Hosting, monitoring and updates with agreed response times.

05 — Data & control

You know where your data is processed. Before we start.

Which AI model we use depends on how sensitive your data is. There are three options.

  1. 01

    Open model, run in Switzerland

    An open AI model running on servers in Switzerland.

    Suited to Sensitive data that should not leave Switzerland.

  2. 02

    Leading model under a data processing agreement

    A powerful commercial model, with contractual safeguards.

    Suited to Tasks that need the strongest models.

  3. 03

    On your own infrastructure

    The solution runs entirely in your environment.

    Suited to Data that must not leave your organisation at all.

  • Where mistakes would be costly, a person approves.
  • Suggestions and answers are traceable, with a source or document.
  • Whether a provider stores data or uses it for training is clarified and documented before we start.

06 — Team

You work directly with the founders.

No account managers, no handing you off to strangers. The people who scope your project build it and support it afterwards.

  • Saverio Reusser

    Chief Executive Officer

    Sales, client relationships and finance.

  • Kris Müller

    Chief Operating Officer

    Product, delivery and data.

  • Our CTO

    Chief Technology Officer

    Architecture, AI engineering and security.

Which process costs you the most time?

Tell us about it in two or three sentences. We will reply with a suggested time for a free 60-minute discovery call.