Skip to content

Work and systems

For companies participating in tendersprzetargi publiczne

Tenders from two registers on one list, arranged by date.

Tender alerts flood the team with announcements. The system downloads them from the Public Procurement Bulletin and the EU TED, extracts the most important data and arranges them according to the deadline for submitting offers. The team starts the day with one list of deadlines.

Tested on sample data, without customer data.

BZP and TED notices on one list
67
suspicious texts retained for human review
6/6
for reading one advertisement by AI
€0.0008

Syntalith test results on public announcements. Last checked: October 22, 2026.

The estimator opened dozens of ads to find a few important ones.

Alerts catch advertisements by a wide code. You need to open each one, find the conditions of participation and save the date. Terms from two registers live in two places.

  1. 01

    Timing

    The date of submission of the offer is immediately visible.

  2. 02

    Conditions

    The most important data from the announcement without reading the entire document.

  3. 03

    Source

    Link to the official announcement for each item.

From announcement to list of dates in four steps.

  1. 01

    Downloads advertisements

    From BZP and TED, according to saved queries.

  2. 02

    Extracts data

    Date, ordering party, subject and conditions in one format.

  3. 03

    Stops suspects

    Text that looks like a command to AI is waiting for a human.

  4. 04

    The team chooses

    He assesses the opportunity and decides in which tender he will participate.

What the system does and what stays with the team.

Makes a system

  • Downloads advertisements from two registers
  • He arranges them according to the deadline for submitting offers
  • Provides a link to the official source

The team decides

  • Which tender to participate in
  • What about frozen text?
  • Preparation and submission of an offer

Statistics: how much does manual tender review cost in 2026.

According to the Syntalith test (2026), the system collected 67 announcements from BZP and TED on one list.

Reading one advertisement by AI cost €0.00083 in the test.

All 6 AI command texts have been retained for review.

According to the Central Statistical Office, the average salary in the enterprise sector in September 2026 was PLN 9,266.48 gross. With employer contributions and 168 hours a month, an hour of work costs the company PLN 66.45.

Scenario: 720 ads per month

By hand
720 × 8 min = 96 h of work, PLN 6,379
Reading by AI
720 × 0.00083 PLN = 0.6€0
He stays with people
selection of tenders and preparation of the offer

Estimate. Assumptions: 720 ads per month and 8 minutes of manual review per ad are sample numbers.

What stays with the team.

The decision to start
The system collects material. The team decides about participation in the tender.
Suspicious text is waiting
Content that attempts to issue commands does not affect the list.
Official source
Each entry leads to an announcement in the register.

This is what team work looks like.

Screens from the test on public announcements.

  1. Zacznij od terminów
    01Zacznij od terminówZespół widzi, co i kiedy trzeba złożyć.
  2. Otwórz ogłoszenie
    02Otwórz ogłoszenieKażdy rekord prowadzi do oficjalnego źródła.
  3. Sprawdź zatrzymane teksty
    03Sprawdź zatrzymane tekstyPodejrzana treść nie wpływa na ocenę przetargu.

Does your team also review tenders manually?

Describe your process to us. We will show the prototype on your process and your data. The analysis and prototype cost €0, and you know the fixed implementation price before the contract.

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.