AI automations for business
Every morning someone reads the inbox and routes the cases, someone retypes data from PDFs into the system, someone glues together the same report on Monday: count those hours per week, multiply by the rate, and that is what the status quo costs you. No one minds it, because the rules live in people's heads, not in the system; the automation reads, classifies and moves data within those rules, and only the exceptions still reach people. From 15 000 PLN net per process, 2–6 weeks to production, a fixed quote before the contract, and a pilot on your data has to hit a written target before we roll out the whole thing.
In short
- Entry
- from 15 000 PLN net · 2–6 weeksper process; a fixed quote before the contract
- What leaves your desk
- Manual retyping and case sortingthe system reads, classifies and moves data; only exceptions reach people
- Automation, not an agent
- When rules describe the whole pathwhen the case content picks the path, that's work for an agent
- Not for
- Automating everything at oncewe start with the one process that hurts most
Problem
Someone reads the inbox and decides where each case goes. Someone retypes data from PDFs into the system. Someone stitches a report together every Monday. Hours of work with no decision in them, and the cases queue up anyway. Count those hours per week and multiply by the rate: that is the cost of the status quo, where no one owns these cases and the rules live in someone's head, not in a system.
Outcome
In 2–6 weeks we take over one such process: the system reads, classifies and moves data between your tools, and only the exceptions that need a decision reach people. Those hours go back to the team, and cases stop waiting in the queue. From 15 000 PLN net, with a fixed quote before signing. The pilot on your data has to hit the written target before you commit to the full build.
This is one process, taken over.
From the case arriving to the trace it leaves. The green line is the boundary the system does not cross alone.
- 01
A case arrives
an email, document or form enters the system
- 02
Classification
the system reads its type, priority and owner
- 03
Written to systems
CRM, ERP or a sheet; every change lands in the log
- 04
Confidence gate
each case gets a confidence score; above a threshold we set with you it moves on by itself, below it waits for a human
Handled automatically
a routine case closed, with a trace
Exception to a human
a decision the system does not make alone
We set the boundary where a human takes over during scoping. Operations with production impact can require approval.
Work off your team's plate
When this fits
When one process eats your team hours every week and is still done entirely by hand.
You don't need to know whether you need an automation, an app, or an agent. You bring the process; on the free scan we name the form plainly, against the seven acceptance criteria. If something simpler is enough, we say so.
Scope
- Inbox triage: classification, case records, replies under an agreed policy
- Documents: data extraction from PDFs, invoices and forms straight into your system
- CRM and ERP: record updates and API integrations, with a change log
- Reports and notifications generated automatically by rule, before the meeting
- An exception path to a human and handover documentation for your team
Not for
- Processes with no owner on the company side
- Websites and marketing chatbots
- Automating “everything at once”: we start with one process
The four processes we most often start with.
Data & compliance
We work in an agreed environment where one client's data is separated from another's. Exceptions and sensitive operations go through a human (EU AI Act Art. 14), and every action stays in the log (Art. 12). After launch the system is monitored. Separately: if the automation replies to users itself, we label that they are talking to a system (EU AI Act Art. 50, from 2 August 2026).
Entry price
from 15 000 PLN
Delivery window · 2–6 weeks
The price follows the number of data sources and integrations, not the number of agents.
You get a fixed quote in the proposal after the free scan: a 30-minute call, a written takeaway in 2 business days, no obligations. You know the full cost before signing, with no hidden items.
What drives the price
- The number of data sources and integrations
- The complexity of the rules and exceptions
- Security, consent and approval requirements
Cost over time
- Implementation buildfrom 15 000 PLN2–6 weeks
- Maintenance and oversightpriced individuallymonthly
Free process scan
Start with a free process scan.
- 30 minutes with the engineer who would build it, not a salesperson.
- A review of the processes that cost you the most time and money.
- A written summary: what to automate, in what order, with cost ranges.
No sales deck and no obligations. If automation doesn't make sense, we'll write that too.
0 PLN
30 minutes · written takeaway within 2 business days
You leave with a plan, not a sales pitch.
- 30 minutes with the engineer who would build it, not a salesperson.
- A review of the processes that cost you the most time and money.
- A written summary: what to automate, in what order, with cost ranges.
No sales deck and no obligations. If automation doesn't make sense, we'll write that too.
Automation questions
What if something simpler than AI is enough?
Then that's what we recommend. If a script or configuring tools you already own does the job, we'll say so in the scan takeaway. That project is cheaper and faster, and an honest answer is a result too.How long does the build take?
Usually 2–6 weeks from contract to production. The timeline is set in the proposal, not mid-work. You work with the engineer who builds it, not a salesperson.How much does AI automation cost?
From 15 000 PLN net for one process. The price follows the number of data sources and integrations, not the number of “agents”. You get a fixed quote in the proposal after the free process scan: you know the full cost before signing.What access does the automation get to our systems?
The minimum it needs: specific actions under specific conditions, with every change logged. We treat the content of emails and documents as data to process, not as commands: even if someone writes instructions into an email or invoice, the automation only performs the actions defined in its scope. Operations with production impact can require human approval: we set that during scoping.What about GDPR and our data?
We work in an agreed environment where one client's data is separated from another's; a dedicated environment is available on request. The automation runs on minimum permissions and logs every operation. We model our discipline on the trace, human-oversight and monitoring regime the EU AI Act requires of high-risk systems, regardless of how your system is classified. Boundaries and approvals are set with you during scoping.What if the automation gets something wrong in production?
Doubtful cases do not move on by themselves: they wait for a human decision, and an operator can stop the system at any time. Every action is in the log, so an error can be reconstructed and rolled back. After launch the system runs under maintenance, so fixes have an owner.