Skip to content
SSyntalith
Book a free process review (30 min)

We implement AI where the return can be measured.

Before you pay, we test AI on your process and your data.

We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓

How we workDetails ↓

  1. 01Call30 minutes with an engineer, free
  2. 02PrototypeA working system on your process and data, free
  3. 03ImplementationTarget and fixed price in the contract; 50% at signing
  4. 04Acceptance50% once the agreed target is met

The agreed target and 50/50 payment apply to automation and custom tools.

Example: one month of invoices
Agent · ready to work

Client results.

All projects →
foap.comcustomer service

foap.com customer service: 95% of replies are sent automatically.

Users from several countries write to foap.com, in several languages. The system reads the entire conversation and the terms of that person, and then sends a response in that person's language. Questions about money and exceptions go to the team with a ready draft, so they have to check and send them.

the system sends the answers itself, without human typing
95%
time from the arrival of the e-mail to the ready response (median)
30 s
e-mails that require a response receive a ready-made draft
100%
less work for every 1000 answers (estimate)
47 h

Data from foap.com production, October 20-21, 2026. Estimate: 3 minutes for handwritten response.

How foap.com customer service works →
Sample messages. Money matters always come to the person.

Other implementations

jaktako.design

studio projektowe

−87%

Time to look for new customers, every day

4 hours → 30 min

approx. 73 hours less per month (21 days of work)

according to the studio

How the agent loop works →
Signature Estates

real estate

approx. −90%

Operator time spent browsing portals, per day

2–3 hours → 20 min

approx. 45 hours less per month (21 days of work)

respect · contact the seller 5 minutes after the announcement

How the agent finds property listings →
Kuba Koziej, CEO and co-founder of MoreGrowth and board member at Natu.Care

Kuba Koziej

CEO and co-founder, MoreGrowth

Board member at Natu.Care

MoreGrowth

„We started with my own AI agent. During the training that followed, I learned to build new applications myself. Today I am bringing this way of working with AI into my business. Working with Syntalith gave me exactly the skills and tools I needed.”

View the original training review on LinkedIn ↗

What our OpenAI and Anthropic partnerships mean for you.

OpenAI makes ChatGPT and Anthropic makes Claude. We belong to the partnership programs of both companies, so we work with their tools every day and learn from the creators themselves.

OpenAI Select PartnerClaude Partner Network
01We will select a tool for the task
We implement both ChatGPT and Claude. We check both on your documents, and you choose the one that makes fewer corrections and saves more time.
02Knowledge straight from the creators
We learn from official OpenAI and Anthropic courses and materials. Your team learns about the features and settings that work today.
03Licenses without our margin
When we roll out ChatGPT, Claude or Copilot, you buy the licences directly from OpenAI, Anthropic or Microsoft and they invoice you. We add nothing to the licence.
04Certified engineers
On the team, we have an AWS certification in machine learning and an Anthropic certification in systems architecture on Claude.

Participation in programs does not mean that OpenAI or Anthropic endorses our services.

Six questions we hear before every implementation.

We answer each with a number with a source.

  1. 01„I don't need it.”
  2. 02„I was checking the AI. It is expensive and weak.”
  3. 03„I'll wait some more.”
  4. 04„I already use AI.”
  5. 05„Is my data safe?”
  6. 06„Why should I trust you?”

„I don't need it.”

Often you need. Sometimes not. We check it before you spend a penny.

How much time AI saves in specific tasks

According to Anthropic (2025), AI shortens the writing of invoices, notes and documents by 87%, and the compilation of information from reports by approximately 95%.

This is an estimate from 100,000 conversations with Claude. They don't take time for corrections outside of the conversation. Studies in companies with a control group showed less, from 14% to 56%, and sometimes work with AI took longer. That's why we measure on your matters.

Source: Anthropic, Estimating AI productivity gains (November 2025) ↗

What have people who work with AI gained?

According to the Anthropic study (2026), among approximately 9,700 people who work with Claude, 86% work faster and 27% save on services that they previously had to purchase.

Source: Anthropic Economic Index, June 2026 ↗

At our client: jaktako.design is looking for new clients 30 minutes a day instead of 4 hours. See how it works →

When AI doesn't make sense

projects with AI agents will be interrupted by the end of 2027: due to costs, unclear profit or lack of risk control
40%+
projects with AI agents will be interrupted by the end of 2027: due to costs, unclear profit or lack of risk control
of the thousands of companies selling "AI agents" have real agents. The rest Gartner calls agent washing: a regular chatbot with a new name
approx. 130
of the thousands of companies selling "AI agents" have real agents. The rest Gartner calls agent washing: a regular chatbot with a new name

Source: Gartner, June 2025 ↗

Typical design

  1. idea
  2. expensive prototype
  3. lack of purpose
  4. abandoned

At our place

  1. the cost of today's work
  2. prototype €0 on your matters
  3. purpose in the contract
  4. system at work
  1. 01

    First we count

    How much does this job cost today. When the profit is too small, we write it clearly and recommend a simpler solution.

  2. 02

    Agent only where needed

    Often, simple automation or a well-set ChatGPT is enough. Then we recommend it.

  3. 03

    Prototype in your process

    20-100 cases that the team has already handled and a goal set before the start.

  4. 04

    You pay for the result

    You pay the second installment only when the system reaches the goal specified in the contract.

You don't know if it will be worth it for you? We will check it with a free prototype on your trial.Schedule a free analysis →

„I was checking the AI. It is expensive and weak.”

A year ago, AI matched experts in every other task. Today it's 85%.

What is GDPval test

OpenAI collected 1,320 real tasks from 44 professions, prepared by specialists with approximately 14 years of experience: legal writing, technical drawing, conversation with a client, nursing care plan. One task took an experienced specialist on average about 7 hours. The same task was performed by a human and an AI, and experts from a given profession evaluated both versions without knowing which was which.

In how many tasks did the AI perform as well as or better than the expert?

According to OpenAI (2026), GPT-5.5 performed 84.9% of GDPval tasks as well as the expert or better; in September 2025 the best result was 47.6%.

Source: OpenAI, GDPval ↗ · GDPval, scientific publication ↗ · OpenAI, GPT-5.5 (April 2026) ↗

And it's cheap. Same job: human and AI, October 2026

Read the 40-page contract and list the risks

EmployeePLN 199 · 3 hours
AI€0.67 · 2 min · Claude Opus 5.5

298× cheaper

Describe and assign 100 cost invoices

EmployeePLN 332 · 5 hours
AIPLN 2.72 · 10 min · Claude Sonnet 5.5

122× cheaper

Reply to 50 emails from customers

EmployeePLN 332 · 5 hours
AIPLN 2.14 · 5 min · Claude Sonnet 5.5

155× cheaper

Scenario. Working hour: approx. PLN 66, i.e. average gross salary according to the Central Statistical Office (PLN 9,266.48, September 2026) with employer's contributions, 168 hours a month. Employee time is our assumption. The cost of AI from the model price list, with the number of tokens typical for such a document, NBP exchange rate PLN 3.8889 per USD 1. The AI result is always checked by a human.

Source: Central Statistical Office, salaries September 2026 ↗ · Anthropic price list ↗ · NBP exchange rate ↗

How we make sure AI is cheap

Every AI company has several models: powerful and expensive, or fast and cheap. The model may also think longer or shorter on a task, and longer thinking costs more. A cheap model is enough to sort emails. We take the stronger one to analyze the contract.

How much does one difficult test task cost with different settings

Points are the result of a set of difficult tests in knowledge, reasoning and programming (Artificial Analysis, October 2026). The more points, the better the model copes with difficult tasks.

The cheapest setup costs 75 times less than the most expensive one and scores fewer points. Therefore, we check in your cases which setting is sufficient. You don't pay for stronger ones without a reason.

Which lowers the bill even more

  • We don't pay for the same thing twice. If the AI reads the same instructions and documents for each task, Claude charges 10% of the usual price for this repetition.
  • Non-urgent tasks go in bulk. A report that is due in the morning may have to wait several hours. In OpenAI it costs half as much.
  • Spending limit. The account has a monthly limit and sends you an alert before you exceed it.
  • Test before each change. We check the new model based on your concerns: whether it works as well and how much it costs.

Source: Artificial Analysis, model ranking ↗ · Anthropic price list (cache) ↗ · OpenAI Pricing (Batch) ↗ · NBP exchange rate ↗

We will count the cost of AI on your cases before you pay a penny.See how a prototype is created →

„I'll wait some more.”

Your competition is already implementing it. And he gets money for it.

Congratulations. You've lived to see the times when AI is cheap and on par with experts. Your competition has it too. And he's already implementing it.

„Internet? Nonsense.”

Newsweek, 1995

This is what was once written about the Internet. It's hard to find companies that didn't believe in him today.

Clifford Stoll, "The Internet? Bah!" ↗

42%

Polish companies are already using AI

3 times more than a year and a half ago

Every third company has not started yet. For now.

Q2 2026 · NBP

Many today say that the AI bubble will burst. The Internet bubble burst in 2000, and the Internet stayed with us forever. It will be the same with AI.

The professions are already changing. According to the Stanford AI Index (2026), employment of programmers aged 22⁠–⁠25 in the US has declined by almost 20% since 2024.

Don't be surprised when a competitor does 2-3 times more work with the same team (this is an example).

Source: Stanford HAI, AI Index Report 2026 ↗

Polish business, measured

InPost runs on Claude across the company. Its founder estimates €70m of extra EBITDA in 2026.

of extra EBITDA in 2026 by Rafał Brzoska's estimate. At least twice that in 2027.
€70m
Claude Code licences at InPost, company-wide since February 2026.
2,200
the work Rafał Brzoska gets through as a manager compared with January, by his own estimate.
2×

Since February 2026 InPost has worked in Anthropic's Claude Code across the whole company. The teams build the tools themselves: foremen for route planning, finance for fraud prevention. Teams compete to show new uses.

„This year alone we estimate it will give us 70 million euros of extra EBITDA. […] Next year it will be at least twice that.”

Rafał Brzoska, founder of InPost (translated from Polish)

„Polska ma Moc” conference, recording published on 15 October 2026.

Recording of the talk, Biznes Info (2026) ↗

Figures checked on 25 October 2026.

PLN 540 million

in grants is waiting for companies this autumn in the SMART Path and Dig.IT

Co-financing for companies, October 2026

Check the conditions of each recruitment in the announcement. The price of our implementation is fixed, with or without a subsidy.

Source: NBP, Quick Monitoring (July 2026) ↗ · Central Statistical Office, Information society in Poland in 2026. ↗

You don't have to wait for the subsidy to count the profit. The analysis and prototype cost €0.Schedule a free analysis →

„I already use AI.”

Your people too. Just no rules.

From our conversations with companies

5 na 10 employees paste sensitive data into private AI accounts: contracts, invoices, customer data.

In ESET survey (2026)

1 na 10 admits this in the survey.

Customer data pasted into a private AI account without an data processing agreement is a violation of the GDPR.

Don't worry, we won't tell anyone :)

What do we see in companies that "already use"

  • Only ChatGPT or Claude accounts, without plugins, skills, integrations and MCP servers with company tools.
  • Company programs are not connected to AI. People still click through systems and Excel, and enter chat data manually.
  • Lack of company standards: everyone writes instructions in their own way, no one shares what works.
  • No rules for personal data, GDPR and company secrets.
  • No limits and no visibility: no one knows who uses AI, how much and for what.

62% Polish employees use AI at work · 27% companies have written rules for working with AI.

Source: ESET and DAGMA, Cyberportrait of Polish business 2026 (March 2026, 1,026 people) ↗

ChatGPT or Claude space review

Accounts, permissions, rules for working with data and 5 tasks set for the team. 1 day.

€450 net

ChatGPT rollout →Claude rollout →

Company accounts, rules of working with data and team training in 2 weeks, from €1,150 net.View implementation for the team →

„Is my data safe?”

Yes. You choose where they go: from your company account to the model on your server.

  1. 00Employee's private account

    Customer and company data ends up in an account that the company cannot see or control. There is no data processing agreement, so it is a violation of the GDPR. When an employee leaves, the chat history goes with them.

  2. 01ChatGPT or Claude corporate account

    The provider does not use your data to train models. A copy of the queries remains in his logs for up to 30 days.

  3. 02API with zero retention (ZDR)

    The provider processes the request and does not save anything. It requires his consent, we prepare the application together.

  4. 03Processing in the EU

    Queries are counted and stored in Europe: OpenAI in the EU region, Claude via clouds with the EU region.

  5. 04Model at your place

    The open model runs on your server or in your cloud in the EU. The data does not go outside the company.

What we do with every implementation

  • Confidentiality agreement and data processing agreement before we receive any data.
  • You choose the cases for the prototype and we anonymize them upon request.
  • The data stays in the EU and we delete it after the test.
  • Corporate accounts with an administrator on your side.

When data cannot leave the company

The open model runs on your server or in your cloud in the EU. To choose from:

White-tailed eagle
Polish model, SpeakLeash and ACK Cyfronet AGH
PLLUM
Polish model, HIVE consortium commissioned by the Ministry of Digitization
Qwen
Alibaba
GLM
Zhipu AI
DeepSeek
DeepSeek

We select a model for the task and check it on your cases. We run it, tune it and maintain it.

Source: Bielik and PLLuM, ACK Cyfronet AGH ↗ · PLLuM, Ministry of Digitization ↗

Do you have a lot of your own data? We will fine-tune the small model.

  1. 1 Your examples
  2. 2 tuning
  3. 3 test on new cases
  4. 4 model at your place

When you have thousands of similar emails, documents or reports, we teach a small open model based on your examples. It works for you, responds quickly and costs less to maintain than a large model. We measure quality on issues that the model has not seen while learning.

Przemysław and Piotr did it in large programming and consulting companies. Artem did it before and does it today at Syntalith. It's about models on your own infrastructure, tuning and quality measurement.

Private AI models →

Source: OpenAI, training on data from private accounts ↗ · OpenAI, who sees conversations ↗ · Anthropic, changes for private accounts (August 2025) ↗ · OpenAI, data in API ↗ · AWS Bedrock, EU profile ↗

You choose the protection level during the interview. We sign the data processing agreement before we see any data.Schedule a free analysis →

„Why should I trust you?”

Because the system is built by engineers with over 22 years of experience.

You talk to people who then write code. They have worked in large companies and built systems that operate under high traffic.

You won't spend money on a prototype that won't do anything.

First a free conversation, then a free prototype in your trial. If we don't see a return, we don't build and write it straight. You only pay for a system that makes sense.

Polish-Ukrainian team of engineers.

Artem, Przemek and Piotr studied computer science together at the Lublin University of Technology in 2017⁠–⁠2021. Artem and Serhii founded Syntalith together. They both come from Kiev and have been living in Poland for years. We build systems for Polish companies, in Polish and according to Polish law.

The band's career over the years

2017–2021: Artem, Przemek and Piotr study computer science together at the Lublin University of Technology.

Serhii Ivanchatenko7+ yearsLinkedIn ↗CRM and PLM systems for global companies
Artem Lisovtsov6+ yearsLinkedIn ↗from backend to AI architecture
Przemysław Markiewicz5+ yearsLinkedIn ↗GenAI team technical lead
Piotr Łach4 yearsLinkedIn ↗backend and RAG in high-traffic systems

Start with a team or with one process.

The team works with AI.

ChatGPT or Claude in 2 weeks: accounts, principles of working with data, training and measurement after 30 days.

from €1,150 net

team of up to 10 people

ChatGPT rollout →Claude rollout →

The system does the work for the team.

Automation, AI agent or new tool. First, a free prototype for your process, then a fixed price.

from €3,000 net

50% payable upon delivery

AI automations →AI agents →

Full pricing →

How we build automation and agents

From a conversation to a working system in four steps.

  1. 01 · 30 min

    Free process review

    €0

    You're talking to an engineer. You receive the written plan within 2 business days.

  2. 02 · 5 days

    Free prototype

    €0

    The system works on 20-100 cases that your team has already handled. You see every result.

  3. 03 · 2–16 weeks

    Build

    50%

    Fixed price and goal written in the contract. First installment upon signing.

  4. 04 · after reaching the goal

    Reception

    50%

    You pay the second installment when the system reaches its target on your cases.

How a free prototype is created

Mon.

We set a goal

What result do you consider good?

Tue–Wed

The system is working

It goes through your affairs, one by one.

Thu

We compare

We check every difference with the team's work.

Fri

Result and price

List of errors and fixed implementation price.

You choose the matters. We anonymize them upon request, they stay in the EU and we delete them after the test.

On the market, such a test usually costs from PLN 3,000 to PLN 8,000.

Book a free process review (30 min)

We establish a prototype during a free analysis.

Data checked on October 25, 2026.

Let's talk

Just write, call or choose a date.

We work fast because we hate corporate cotton wool. We don't have salespeople or managers to keep an eye on managers. You talk immediately to engineers who then build your system. We respect each other's time and yours too.

We will accept 3 more companies by the end of the year.

Write
contact@syntalith.ai
Call
+48 888 78 48 78
WhatsApp
write on WhatsApp

Mon–Fri 9:00–17:00 CET

Common questions

  • How much does an AI implementation cost?

    ChatGPT or Claude for a team of up to 10 people costs from €1,150 net. Automation or an AI agent costs from €3,000 net. Testing the process on 20-100 cases your team has already handled is free. We write the fixed price in the contract before we start.

  • Can automation pay off for a smaller company?

    Use your own figures: hours per week × full hourly labour cost × 52 weeks. If the annual process cost is lower than the implementation price, we will say so in the review and recommend a simpler solution. Sometimes a feature in a tool you already have is enough.

  • Do we need to change the tools we use?

    Usually not. We connect to the tools you already use through an API, file imports and exports, or your inbox. We check the connection during the review, before quoting.

  • Our employees already use ChatGPT. Isn’t that enough?

    For writing and summarizing, this is often enough. The implementation organizes accounts, data rules and projects for tasks that come back every week. Once you have the license, we start with an inspection of the space in 1 day for €450 net.

  • What does the free process review commit us to?

    You get a 30-minute call with an engineer and a written summary within 2 business days. If we do not see a worthwhile next step, we say so in the document. The decision stays with you.

  • What if AI makes a mistake?

    The system itself does what it is sure of. Lack of data, conflicting sources or an issue outside the rules are sent to the person with a short summary. Every decision leaves a trace. We set the acceptance threshold before construction, and you pay the second installment after acceptance.

  • Where is our data processed?

    Before starting, we sign a confidentiality agreement and an data processing agreement. You choose the cases for the test, we anonymize them upon request, they stay in the EU and we delete them after the test. In ChatGPT and Claude enterprise plans, the provider does not train models on company data by default.

What our OpenAI and Anthropic partnerships mean for you.

OpenAI makes ChatGPT and Anthropic makes Claude. We belong to the partnership programs of both companies, so we work with their tools every day and learn from the creators themselves.

OpenAI Select PartnerClaude Partner Network
01We will select a tool for the task
We implement both ChatGPT and Claude. We check both on your documents, and you choose the one that makes fewer corrections and saves more time.
02Knowledge straight from the creators
We learn from official OpenAI and Anthropic courses and materials. Your team learns about the features and settings that work today.
03Licenses without our margin
When we roll out ChatGPT, Claude or Copilot, you buy the licences directly from OpenAI, Anthropic or Microsoft and they invoice you. We add nothing to the licence.
04Certified engineers
On the team, we have an AWS certification in machine learning and an Anthropic certification in systems architecture on Claude.

Participation in programs does not mean that OpenAI or Anthropic endorses our services.