Skip to content
VistulaCode

MEET THE FOUNDER

Hi, I'm Michał.

I'm Michał Araszkiewicz, founder of VistulaCode. I help teams make more time for customers and work that needs their expertise. You work directly with me to assess the opportunity, test a tool and bring it into daily use. Based in Poland, I focus on product listings, order entry and customer replies for businesses with 30–500 employees.

Let's talk about your businessHow we work
Michał Araszkiewicz, founder of VistulaCode
Michał AraszkiewiczFounder of VistulaCode

MY BACKGROUND

Michał Araszkiewicz

Founder of VistulaCode · Head of Product for VirtualCar360

I run VistulaCode and work as Head of Product for VirtualCar360. In my automotive software work, I define how tools should work for their users, check the quality of AI-edited images, test data exchanged between systems and prepare instructions for staff and developers.

Let's talk about your business

EXPERIENCE RELEVANT TO YOUR PROJECT

Start with the work.
Build around the people doing it.

My product work starts with what people need to get done and how we'll know a tool helps. At VistulaCode, I apply that approach to product listings, orders and customer replies: agree the useful outcome, then test it on real examples.

  • 01

    Defining what the tool should do

    Turning everyday work into a list of tasks, screens and rules that developers can build.

  • 02

    Checking AI results

    Agreeing what a good result looks like, finding errors and testing changes on realistic examples.

  • 03

    Connecting business systems

    Defining which information moves between systems, then checking the result and what happens when something goes wrong.

  • 04

    Preparing staff and developers

    Preparing clickable test versions, guidance on connecting systems and instructions for new users.

SELECTED PROFESSIONAL WORK

The experience behind the approach.

Selected examples from my professional work in automotive software, showing what I did and whether the work reached everyday use, research or a test version.

01

In everyday use

Sending the right vehicle photos for processing

  1. Select vehicle session

  2. Process its photos

  3. Check status and previews

My involvement
I defined how the software should find the correct vehicle photo session, copy its images, send them for processing and show the result. I also defined the checks needed to confirm it worked.
What the work produced
The working software included a panel where authorised staff could see processing progress and photo previews. My testing also found a case where the wrong photo session was selected, so it could be corrected.

Why it matters for your projectWhen information moves between systems, staff need to know that the right record was selected and to see what needs checking.

02

Research and testing

Checking the quality of AI image processing

  1. Define image quality

  2. Review AI image errors

  3. Test proposed changes

My involvement
I set quality requirements for vehicle background removal and replacement, including how windows, reflections and bodywork should look, and directed evaluation of proposed changes.
What the work produced
The work produced ways to measure image quality, a review of the data and checks of the resulting images. The proposed improvements were still being tested and were not approved for everyday use.

Why it matters for your projectBefore spending on setup, your team needs realistic examples and an agreed way to decide whether the test worked.

03

Clickable test version

A clearer video request process

  1. Choose video format

  2. Select video scenario

  3. Track simulated request

My involvement
I defined and revised the steps for requesting vehicle videos: choose a desktop or mobile format, select a video scenario, submit the request and follow its progress.
What the work produced
A clickable test version, guidance on connecting systems and instructions for developers. Video generation was simulated; connecting a live video service was a separate step.

Why it matters for your projectA clickable version lets the manager and developers agree how the tool should work before paying for the full build.

These examples describe my professional background, not client projects delivered by VistulaCode. VistulaCode focuses on work in wholesale, distribution and e-commerce. It does not offer automotive imaging or vehicle media projects.

WHAT WE HELP YOU CHANGE

Give your team time back.
Less time moving data around.

A supplier sends a spreadsheet that needs cleaning before products can go online. A customer emails an order that someone retypes into the order system. A support agent searches three systems to answer a delivery question.

We work with the people who do the job to identify where a change would help. Then we agree what AI prepares, what staff check and what happens when information is missing.

The audit gives you a plan, the paid pilot tests the benefit, and implementation puts the approved tool into daily use. You decide on each purchase separately.

See the three tasks we help with
AI-generated illustration of Michał explaining a business workflow during a workshopAI AT WORK

Example: answering a delivery question

  1. Customer asks about delivery
  2. AI drafts a reply with references
  3. Staff check and send

Give staff the suggested reply and order details together, so they can check the answer instead of searching several systems.

AI-generated illustration

AT A GLANCE

Who you'll work with
and what to expect.

The essentials when you are choosing who to work with.

Based in PolandWorking with companies worldwide, online
What VistulaCode is
An independent business helping teams reduce repetitive work on product listings, orders and customer replies with AI tools.
Based in
Poland (Central European Time)
Delivery
Working online with companies worldwide
Working language
English
Typical clients
Wholesalers, distributors and e-commerce businesses with 30–500 employees
Services
Product Catalog Automation; Order Processing Automation; Customer Support AI
What you can buy
Automation Audit, a separately priced AI Pilot, Implementation & Team Training, and optional Monthly Support
Pricing
Audit: $1,250 USD net. Order-entry pilot example: $3,000–$5,000 USD net, with a fixed quote before work starts. Implementation and training: from $6,500 USD net, excluding audit and pilot. Optional support: from $1,000 USD net/month. AI provider charges, software subscriptions and hosting cost extra.
Out of scope
We don't build general-purpose AI models from the beginning and don't take on automotive imaging or vehicle media projects

WHAT YOU CAN CHECK

An investment you
can explain to your colleagues.

A budget decision needs more than a description of the technology. Each stage gives you written findings, costs and evidence to discuss with the people approving the investment.

  • Work and price in writing

    Before each stage, you receive a list of what we will build or review, its price, who does what and how the result will be checked. Extra tasks or system connections need a separate agreement.

  • Value measured on your work

    We measure the task today, then compare it with using AI, including staff review and corrections. Running costs are part of the comparison, so you can assess the full picture.

  • Your team reviews

    The manager responsible for the task helps test everyday examples and difficult cases. We agree which results need staff approval before they enter another system.

  • Instructions and support

    Before launch, we agree who runs the tool, who fixes problems and which instructions your team receives. Monthly support is optional and priced separately.

WHAT A GOOD PROJECT NEEDS

Real examples.
Shared responsibility.

The service examples show what we could build; the calculator uses assumptions you can change. Your actual benefit depends on the files you provide, how often the task happens, the software you use and the paid test results.

See the approval steps
  • A person responsible for the task who can explain it and check the results.
  • Examples of everyday work and difficult cases, with permission to use them in the review and test.
  • Confirmation from your IT contact that the required records can be read and the agreed systems connected.
  • An agreed measure of success and your written decision before each new paid stage.

LET'S TALK ABOUT YOUR BUSINESS

What would your team do with more time?

In a 20-minute call, tell us what your team types, copies or searches for every day, and what those hours could be used for instead. We'll explore where AI could help and what we need to check.

Let's talk about your business

20 minutes to explore what takes up your team's time