Private AI.In your company,on your terms.
Enterprise grade. Without the enterprise price tag.
We design and deploy private AI solutions for organisations that need measurable results, control over their data and an architecture ready to maintain.
- An agreed start date
- A quote once the scope is defined
- GDPR / AI Act in scope
Artur · software architect · 18 years of experience
Meet the person accountable for deliveryArtCode Software
We choose how to run the system based on your data and requirements. We check processing regions, prompt retention, use of content for training and vendor access. The terms depend on the service, configuration and contract. We confirm them before choosing a solution.
2.2 · What we agree before implementation
- The level of environment isolation and operator access. A managed service does not always provide separate infrastructure.
- Whether the vendor retains prompts and outputs, for how long, which exceptions apply and whether content is used for training.
- Whether vendor staff can access content, in which situations and on what basis.
- What logs contain, how long they are kept and how prompt content is limited in logs.
- Who controls keys, permissions and logs, and how access is handed over when the engagement ends.
- Processing regions, sub-processors and transfers outside the EEA. We check GDPR and AI Act requirements relevant to the use case.
2.3 · Vendor selection and negotiation
We pick the provider of the server, GPU server or endpoint and negotiate the terms: processing region, no training on your content, retention, sub-processing, DPA and SLA.
Security requirements go into the order, not into a slide deck.
Demonstration · fictional data
From document to answer
01 / Document
Document submission: Tuesday by 12:00. Approver: operations manager.
When must the documents be submitted?
This example demonstrates the mechanism. It does not validate performance on your company's documents.
First, we establish whether AI makes sense for your business.
Does AI make sense for your process?
60 minutes with Artur. We look at one process and establish what is worth testing next.
- An assessment of where AI could help in the described process
- Key constraints and missing information
- A short written summary of conclusions and the next step within 3 business days of the call
PLN 299 total
60 minutes online · short summary included
The consultation fee is credited against an audit ordered within 30 days of our meeting.
Choose a time and payWe discuss one process based on your description. The consultation does not include additional research, an audit or architecture design. It does not commit you to implementation; postponing investment may be the recommendation.
Service and payment information
Want to check the fit first? A short introduction, up to 15 minutes
We can establish whether I can help and which scope of work is worth considering, free of charge. Process analysis and recommendations belong to the paid consultation. Describe your situation in three sentences.
AI Readiness Audit
The entry point. You know what makes sense, what it costs and where to start.
- Current state and the gap
- Risk, compliance, priorities
- Recommendation and a pilot plan
The price depends on how many processes there are and how complex they get. Two simple processes cost PLN 4,900; six complex ones cost more. You know the scope and the amount before the start.
If you move into a pilot after the audit, the audit fee is deducted from the pilot price.
Audit, pilot and support prices are net and indicative. The cost of servers, AI endpoints and tokens sits with the client and is not included in those amounts. We confirm the exact project price once the scope is defined. The consultation price is stated separately including tax.
See the scope, pilot and support costs
Net amounts
Before work starts, we confirm the scope, price and timeline. The consultation helps establish which next step makes sense.
Annex 1
| No. | Item | Net amount |
|---|---|---|
| 1.1 | AI Readiness Audit 2 weeks · fixed scope | from PLN 4,900 |
| 1.2 | Pilot 6–8 weeks · real data | from PLN 19,000 |
| 1.3 | Support and monitoring Production, monitoring and further work. Two tiers, a minimum term of 3 months, notice of 1 month. | from PLN 1,900 / month |
1.1 · AI Readiness Audit
The entry point. You know what makes sense, what it costs and where to start.
- Current state and the gap
- Risk, compliance, priorities
- Recommendation and a pilot plan
The price depends on how many processes there are and how complex they get. Two simple processes cost PLN 4,900; six complex ones cost more. You know the scope and the amount before the start.
The EU AI Act compliance review is part of the package, at no extra charge.
If you move into a pilot after the audit, the audit fee is deducted from the pilot price.
1.2 · Pilot
A working solution in one process, with the success criterion agreed before the start.
- An application environment at a provider we choose together: your server or a data center in Poland or the EU
- Integration with one system
- Measured effect before the decision to scale
- A month of hypercare after handover, included
The price rises with more than one integration, with special category data that carries extra audit requirements, or with your own model on a dedicated GPU machine.
This is an application environment. The hardware layer (servers, GPUs, colocation) and the endpoints and tokens are quoted separately and are not part of this amount.
An earlier audit is deducted in full.
1.3 · Support and monitoring
Production, monitoring and further work. Two tiers, a minimum term of 3 months, notice of 1 month.
| Entry | Flagship | |
|---|---|---|
| net | PLN 1,900 / month | PLN 3,900 / month |
| 1.3.1 | For smaller deployments and companies of 15–30 people. | For production systems a team works with every day. |
| 1.3.2 |
|
|
| 1.3.3 | Hours beyond the pool: PLN 350 / h | Hours beyond the pool: PLN 300 / h |
The cost of models, endpoints and infrastructure sits with the client. You pay for the care, not for the server.
Your own model on a dedicated GPU machine is maintained on the Flagship tier only, after a load analysis and with a separate quote.
Audit, pilot and support prices are net and indicative. The cost of servers, AI endpoints and tokens sits with the client and is not included in those amounts. We confirm the exact project price once the scope is defined. The consultation price is stated separately including tax.
The party signing this document is a person, not a department.
- 8.1
Who signs.
Artur Kluz. ArtCode Software is a sole proprietorship: the architect who designs your deployment is the same person who runs it and the same person who answers the phone when something stops working. 18 years across FinTech, insurance and DMS/ECM systems. The professional profile is public and linked at the signature.
- 8.2
What I am personally accountable for.
The architecture, where your data physically sits, and the acceptance criterion being written down before the work begins. The guarantees in clause 2.2 are contract terms, not statements on a website.
- 8.3
What I will not do.
I will not promise an outcome I cannot write into a contract. I will not start building before the problem is described and the test for “solved” is agreed. I will not put hardware or tokens on your invoice: you choose and pay for the infrastructure.
Explore the details
Open what matters to your decision: compliance, delivery and answers to your questions.
Subject
From framing the problem and auditing readiness through to architecture, a pilot and production maintenance.
- 1.1
AI for documents and knowledge
RAG, OCR and semantic search for processes where knowledge has to be available, reliable and accountable.
- 1.2
Process automation
We design agent workflows and integrations that shorten handling time without giving up control over decisions.
- 1.3
Private AI
On-premise, hybrid, or on a managed endpoint in a PL or EU region. You choose and pay for the infrastructure; we design it, negotiate the terms with the provider and maintain it.
1.4.1
Data security
Your servers or a dedicated environment in PL / EU.
1.4.2
Enterprise quality
Standards and documentation as in a large organisation.
1.4.3
Predictable cost
You know the scope and cost before work starts.
1.4.4
Smooth delivery
We agree on a start date together, once the scope is defined.
Comparison
These are different ways to use AI. Compare the specific plan, contract and configuration, especially when working with company data.
- Where the data sits
Ready-made AI assistantRegion depends on the plan and vendor terms
Automation agencyDepends on the infrastructure and agency contract
ArtCodeWe agree on the region and environment before implementation
- Who holds the keys and the logs
Ready-made AI assistantLevel of control depends on the plan
Automation agencyDepends on the contract and account ownership
ArtCodeOwnership and access are defined in the contract
- What happens to your prompts
Ready-made AI assistantDepends on the plan, settings and service terms
Automation agencyDepends on accounts and tools used
ArtCodeWe confirm retention and exceptions for the chosen service
- Knowledge base built from your documents
Ready-made AI assistantFeatures and permissions depend on the plan
Automation agencyDepends on the chosen solution
ArtCodeYes, held by you, with per-user permissions
- Integration with your systems
Ready-made AI assistantAvailable integrations depend on the plan
Automation agencyReady-made or designed for the process
ArtCodeDesigned around the process, one in the pilot
- EU AI Act and GDPR
Ready-made AI assistantOn your side
Automation agencyCheck the scope of the agency contract
ArtCodeWe agree on the review scope for your use case
- Who maintains it a year from now
Ready-made AI assistantThe vendor maintains the service; you manage your account
Automation agencyDepends on the contract and account ownership
ArtCodeYou or us; the code and the configuration are yours
- Who it is for
Ready-made AI assistantA ready-made assistant for everyday work
Automation agencyAutomation with an agreed scope
ArtCodeA solution designed for your process and data requirements
Compliance
Regulation (EU) 2024/1689
AI Act obligations depend on the use case and your role. We check which apply to your system.
Art. 6
Risk classification
Article 6 and Annexes I and III
Basis: Art. 6 · Classification
Art. 10
Data and its quality
requirements for high-risk systems
Basis: Art. 10 · If applicable
Art. 13
Information for deployers
transparency of high-risk systems
Basis: Art. 13 · If applicable
Art. 50
AI disclosure and content labelling
obligations depend on role and use case
Basis: Art. 50 · If applicable
Art. 72 and 73
Monitoring and serious incidents
obligations for providers of high-risk systems
Basis: Art. 72 and 73 · If applicable
In this format you get a report: not a PDF full of graphics, but a concrete checklist with references to the provision and a list of things to do.
This is a review and a gap analysis against the requirements of the EU AI Act: we name the obligations that apply to you and the gaps left to close. It is not a formal conformity audit and not a certification.
Scope depends on the use case · Regulation (EU) 2024/1689
part of the AI Readiness Audit · from PLN 4,900 · 2 weeks
Working method
The best AI deployments do not start by picking a tool. They start with the process, the data, the risk and the ownership of the outcome. That is why we split the work in two and never mix the parts.
Not agreed
5.1 · The problem problem space
Before anything is built
- Which process hurts, and who runs it today.
- What data actually exists, and in what state.
- What constrains us: law, systems, people, time.
- How we will know the problem is gone.
Output: an analysis report: process, data, risks and constraints, with a one-page summary for the decision-maker.
Scope agreed
The scope agreement is a gate. Press it to see what opens.
We do not enter the solution space until we know exactly what the problem is and how we will know it has been solved.
- 5.3.1
Your data stays under your control.
Your servers or a dedicated environment in a Polish or European data center. Keys and access on your side.
- 5.3.2
Every pilot has a success criterion.
Value, quality and risk are measured before any decision to scale.
- 5.3.3
Architecture built to be maintained.
Integrations, monitoring, ownership and procedures are not an afterthought.
Schedule
- 6.1
A consultation about one process
60 minutes for PLN 299 including tax. We discuss one process, its constraints and whether AI makes sense. You receive a summary with a recommendation within 3 business days after the call.
- 6.2
Plan the next step
If you decide to continue, we agree on the scope of an audit or implementation. We confirm the price, timeline and acceptance criteria before work starts.
- 6.3
Start once the scope is agreed
We agree on a start date after defining the scope and required access. You speak with the person who designs the solution.
Questions
We answer them the same way on this page as across the table: specifically, and without promising things that cannot be kept.
- Isn't this too expensive for a company our size?
- The audit starts from PLN 4,900, and its price is deducted from the pilot. The pilot has a fixed scope and a price you know before it starts, so the risk is limited to its cost. The success criterion is agreed before the pilot begins, so you calculate the return on your own numbers rather than on our promise. You scale only after the effect has been measured.
- What happens to the project if you are unavailable?
- The code, configuration, keys and documentation are on your side from day one. Every deployment ends with a handover of skills to your team or your IT. You can maintain the solution yourself, with us, or with another supplier.
- Where will our data be processed physically?
- We choose how to run the system based on your data and requirements. We check processing regions, prompt retention, use of content for training and vendor access. The terms depend on the service, configuration and contract. We confirm the terms and any exceptions before choosing a solution.
- How much of our team's time will this take?
- The audit: roughly 4–6 hours of conversations with the people who run the process. The pilot: one person on your side as the process owner, around 2 hours a week for decisions and testing. The rest is on us.
- What exactly is in the support price, and what is not?
- In the price: monitoring, updates, service work within the pool of hours, advisory and the report. Outside it: servers, AI endpoints, tokens and licenses. You pay the provider for the infrastructure and us for looking after it.
- Can we have our own model instead of an endpoint?
- Yes, on a dedicated GPU machine in Poland or the EU. By default we recommend a managed endpoint in a PL or EU region with several models and OCR, because it gives better quality and speed at a lower cost. Your own model makes sense for particular isolation requirements and needs the Flagship support tier.
Is your question missing here?
Write it in one sentence and we will answer without an offer attached.