Services — AI automation

Practical AI that takes work off your team.

AI in the processes where it saves time: e-mail, calls, documents. A person approves, and data stays in the EU.

We reply within 24 hours

E-MAIL
FDesk sorts
CALLS
DeskVoice answers
DOCUMENTS
Invoa OCR
CRM
Data entry
PERSON
Checks and sends
EU DATA
GDPR
AI with a person in the loop

SITUATIONS

When AI is worth introducing.

  • 01

    The team is drowning in e-mail

    Every message has to be sorted and answered by hand.

  • 02

    Calls go unanswered

    Outside office hours or at peak times, clients get no answer.

  • 03

    Documents are retyped by hand

    Invoices, contracts and applications are entered into the system manually.

  • 04

    Knowledge lives in people's heads

    Answers to common questions have to be found by asking colleagues.

WHAT WE BUILD

Where AI pays off.

  • 01

    E-mail triage and draft replies

    As in FDesk: AI sorts the message and drafts a reply, and a manager checks it.

  • 02

    Call answering

    As in DeskVoice: AI answers, advises and hands the call to a person.

  • 03

    Document and invoice recognition

    As in Invoa OCR: data from the document reaches the system without retyping.

  • 04

    Assistants built on company knowledge

    Answers from your documents, not from the internet.

  • 05

    Data extraction into CRM

    Information from e-mails and documents lands in the record automatically.

  • 06

    Integrations with existing systems

    AI works inside your tools, not next to them.

HOW WE WORK

Four principles.

  • 01

    One process first

    We start with one process, not with "AI for everything".

  • 02

    We measure the result

    Before and after: time, errors, response speed.

  • 03

    A person approves

    People make the important decisions; AI prepares them.

  • 04

    Data stays in the EU

    Data is processed in the EU, with GDPR requirements built in from the start.

WHERE AI DOES NOT HELP

Honest about the limits.

  • 01

    A process nobody can describe

    If it is unclear how the work should be done, AI will not fix that.

  • 02

    Messy data

    If the information is chaotic, it has to be put in order first.

  • 03

    Decisions without a person

    Where a mistake is costly, a person must approve.

Article: AI Automation for a Small Business: Where It Pays Back and Where It Burns Money →

INTEGRATIONS AND STACK

Models and integrations.

  • E-mail
  • Telephony
  • CRM
  • Accounting
  • EU hosting

We are not tied to one supplier. We choose the model for the task: by quality, price, speed and where the data has to stay. Where needed, we use models that process data in the EU.

HOW WE WORK

From the first call to maintenance.

  1. 01DiscoveryA 30–45 minute call about your process and goal.
  2. 02ProposalScope, timeline and cost in writing.
  3. 03PrototypeA clickable design you approve before any code.
  4. 04DevelopmentA demo every two weeks, a test environment from day one.
  5. 05Launch and maintenanceCode, domain and data stay yours.

SMALL TASKS

A small task is real work too.

A single automation, such as e-mail triage or document recognition: we take it like any project.

Tell us about the task →

OUR PRODUCTS

Our products are the proof.

All products →

FSM PRODUCT · HELPDESK

FDesk

A helpdesk on your own domain. AI sorts every e-mail and drafts a reply; a manager reviews and sends.

Request a demo →

FSM PRODUCT · AI CALLS

DeskVoice

An AI receptionist that answers calls, advises customers from your knowledge base and hands over to a person when needed. Data stays in the EU.

Request a demo →

FSM PRODUCT · E-INVOICING

Invoa

Invoices, UBL e-invoices and accounting in one place: Invoa works directly on top of Maximus.

Request a demo →

FAQ

AI automation questions.

Is my data safe?

Data is processed in the EU in line with GDPR.

How are costs set without a price list?

Cost depends on the process and its scope. We state it in a written proposal after a short call.

Which models do you use?

We are not tied to one supplier. We choose the model for the task: by quality, price, speed and where the data has to stay. Where needed, we use models that process data in the EU.

Can AI make mistakes?

Yes, which is why a person checks the important actions.

What happens after launch?

We maintain the system, measure the result and adjust the process.