Skip to content

Blog · Published · Updated

A custom ATS with a voicebot: 3⁠–⁠4 hours less work per candidate

Short answer

For the HR team of a technology company we built a custom ATS with a voicebot for pre-screening: the recruiter adds the CV and sends a link, the voicebot runs the approved conversation, and the result with its sources goes to the recruiter, who makes the decision. The team works about 3 to 4 hours less on each candidate.

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
net per application, fixed price before contract
from €3,000
from signing the contract to acceptance
4–10 weeks
free prototype on your documents
€0

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

Author: Syntalith · 5 min read

Example: the morning request inbox
Agent · ready

Statistics

Recruitment and AI: statistics 2026

3–4 h

less HR team work per candidate after rolling out a custom ATS with a voicebot for pre-screening at a technology company. Syntalith implementation result, client under NDA.

Figures checked: 22 October 2026

From CV to the recruiter's decision
  1. 01CVthe recruiter adds the candidate
  2. 02Planthe recruiter approves the questions
  3. 03Conversationthe candidate knows they are talking to a voicebot
  4. 04Resulta transcript and sources for every answer
  5. 05Decisionthe recruiter assesses and chooses the next stage

What the HR team gained from candidate screening with a voicebot

  • About 3–4 hours less work per candidate.
  • A simple process: the recruiter adds the CV and sends the candidate a link. The voicebot runs the pre-screening, and the candidate talks when it suits them, while the link is valid.
  • People still run the most important interviews. They come to them with the result of the first conversation, the transcript and the sources.
  • The custom ATS replaced an expensive SaaS system the team used to run recruitment in.
  • Every decision keeps the question plan, the transcript and the version the recruiter accepted. That is the trail you need when a candidate asks what a decision was based on.

Before the conversation: the recruiter approves what to ask

1. Role requirements. The recruiter starts with the job requirements. Every requirement has a source you can go back to.

2. Candidate analysis. The recruiter adds the CV. The system compares it with the requirements and the form, points to areas to check in the conversation and shows the source for each. The AI's interpretation is marked as material for the recruiter to assess.

3. Conversation plan. The recruiter reviews the questions, their order and timing. Every question has a priority, a limit on follow-ups and a recorded reason for asking it. The approved plan is locked for editing, so the candidate hears exactly these questions.

Candidate analysis: areas to check in the conversation, the AI interpretation for the recruiter to assess and the source of the requirement
The candidate analysis points to 3 areas to clarify. Each shows the source of the requirement. Synthetic data. Full screen
The approved conversation plan with seven questions, and the time and priority of each question
The approved plan: 7 questions, about 10 minutes, a 12-minute limit. Every question has a time and a priority. Synthetic data. Full screen

The conversation: the candidate knows who they are talking to

4. Invitation. The candidate gets a one-time link. They install no app and need no camera.

5. The voicebot conversation. The voicebot runs a voice conversation in Polish, with a short part in English. It takes about 10 minutes, 12 at most.

The start screen is labelled as a conversation with an AI assistant. Before the conversation begins, the candidate sees its scope, gives consent and checks the microphone. The screen also says plainly that a recruiter will check the answers, and that the system rejects no candidates. At any moment the candidate can stop the conversation, decline to answer or ask to be contacted by a recruiter.

The candidate's start screen with the conversation time, the format and a note that a recruiter makes the decision
The candidate's screen before the conversation: time, format and a note on the recruiter's decision. Synthetic data. Full screen

After the conversation: a result with sources and a person's decision

6. Result. The system prepares a result with references to the candidate's statements and to the sources. The recruiter checks the transcript, corrects the result and only then approves it.

7. Hand-over. The approved result is passed on through a link with an expiry date. The transcript and the recording have to be ticked separately, and the link can be revoked.

Recordings are encrypted on disk. In the configuration shown, recordings of new conversations and public links expire after 7 days. The person who gets the link sees the approved result and has no access to the recruiter's whole panel.

Candidates who pass the pre-screening go on to interviews with people. The team runs these interviews as before, only now with the result of the first conversation to hand.

Where the role of AI ends

The AI Act lists AI systems used to select and assess candidates in Annex III, among high-risk systems. The obligations for this group will apply from 2 December 2027. The requirement to tell a person they are talking to an AI system has applied since 2 August 2026.

In this project, AI prepares the analysis, runs the approved conversation and puts together a result with sources. The recruiter assesses the candidate and decides on the next stage.

The team's own ATS

The app is the team's ATS and replaced an expensive SaaS system the team used to run recruitment in. Besides screening, it has roles and recruitment stages, notes, tasks, a calendar with export to an ICS file, message drafts and CSV export.

If your team stays with its own ATS, we connect the screening to it. We set the scope of the connection when we price the work.

What a custom ATS with a voicebot costs

An ATS with pre-screening is an app built around the team's process. We build AI apps from €3,000 net, usually in 4–10 weeks. We give a fixed price before the contract, after analysing the process, the data sources and the integrations.

The first step is free: 30 minutes with an engineer about your recruitment. Within 2 working days you get, in writing, which stage is worth moving to a voicebot and the price the rollout starts from.

See the 7 screening steps in the implementation write-up. Each step has a description and a screen you can enlarge.

What candidate pre-screening costs and when it pays back

Scenario: 60 candidates a month
Today
60 candidates × 45 min = 45 h × €34.90 = €1,571
With the system, a person checks the output in 30% of the time
€471
Model, 1 call per case, Claude Sonnet 5.5
€2
Cost of one case
€26.18 today, €7.88 with the system
Saving
€1,098 a month

Payback of €3,000

2.7 months

The volume, the time and the checking share are the scenario's assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat. Model: 8,000 input tokens and 1,500 output tokens per call at Anthropic's rates, converted at the ECB rate of 1.1205 USD per euro.

Next to the client's result, the scenario counts 60 candidates a month and 45 minutes per candidate for the first conversation, the notes and writing up the result.

What we measure in the trial
  • Time from receiving a case to the result, before and after the change
  • Cases the app passed to a person
  • Corrections by the person checking
  • Model and maintenance cost at the target volume

Where to start

In the free process review we choose one process and count the team's volume and working time, and within 2 working days you get a written summary with a price. We build as part of AI apps, and we price maintenance separately.

Worked example · Scenario

How we count the payback: candidate screening

We run the same count for your HR department in the free process review.

€2,327

a month for handling 800 candidate and employee documents by hand

Payback: 1.5 mo.

5 minutes to receive and check a document and checking in 15% of the old time are our assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. Model cost: reading a document in our case file system (about €0.0002) plus a description of what is missing (Claude Sonnet 5.5, about 2,000 + 300 tokens), at the ECB rate of 1.1205 USD per euro on 22 October 2026.

See the full calculation →

This is how we implement it in companies

All case studies →

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

„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 ↗

When an off-the-shelf tool does not fit, we build your own.Book a free process review (30 min) →

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 21, 2026.

„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.

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

Custom AI apps

When an off-the-shelf tool does not fit, we build your own.

We build an application with a language model for the work of your team: searching the company's knowledge, case panel, voicebot or portal. We check the validity of your documents before launch, and the code stays with you.

See AI applications →
net per application, fixed price before contract
from €3,000
from signing the contract to acceptance
4–10 weeks
free prototype on your documents
€0

Questions

  • How does AI help with candidate screening?

    A voicebot runs the first conversation from a plan the recruiter approved, and the result with the transcript and sources goes to the recruiter. A person decides on the next stage.

  • Is AI screening covered by the AI Act?

    Yes. Systems for selecting candidates are in Annex III of the AI Act, and the obligations for them apply from 2 December 2027. Telling a candidate they are talking to AI is required from 2 August 2026.

  • How much does it cost?

    At Syntalith apps, automations and AI agents start from €3,000 net. At 60 candidates a month of 45 minutes each, the payback is 2.7 months.

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.

Mon–Fri 9:00–17:00 CET

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
You buy accounts directly from OpenAI or Anthropic and receive an invoice from them. We do not add anything to the license.
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.