Service · Automation and AI

Every inquiry gets an answer in minutes — at any hour, without your manager

We build AI agents, chatbots and Make scenarios that handle incoming inquiries on their own: calculate quotes, qualify clients, reply on WhatsApp and pass hot leads into your CRM. At night, on weekends, in the client's language. First working scenario — in 1–2 weeks.

See cases

Prices in the open — from €1 000. See prices and what is included ↓

1–2 weeks
To the first working scenario
24/7
Replies to clients at night and on weekends, without people
−54%
Make operations after scenario optimization

Numbers that speak for us

€1.85M
Ad budgets managed through us
8+ years
In marketing and advertising
−28%
Median drop in cost per lead by month two
98%+
Event coverage in analytics
15–20%
Of "wasted" spend we find in the first month
Mzogs Kronitek Edgars Būve Brionix RigaOn Liberta RPI Group
Your situation

This is for you if you recognize yourself

Even one of these points means there's something to automate — and it'll free up your team's time in the very first month.

Inquiries at night and on weekends go unanswered.

A client wrote at 10:30 pm, the manager saw it in the morning — and the client already went to a competitor who replied first.

A manager calculates quotes, copies data and sends emails by hand.

Hours a week go into work that follows clear rules and repeats the same way every time.

Leads get lost between the form, the spreadsheet and the CRM.

An inquiry lands in the inbox but never reaches the CRM; or it lands twice; or with no source — and it's unclear which ad brought it.

Reports are compiled by hand every week.

Someone exports numbers from ad accounts into a spreadsheet instead of working with clients.

The initial back-and-forth eats up your sales team's time.

Half of the incoming isn't your audience, but you still have to sort through it manually before you can tell.

Who it fits

Automation pays off when there's a flow

If you only get a handful of inquiries — we'll say so and won't sell you a scenario for the sake of it.

You get 30+ inquiries a monthand some of them get lost or wait too long for an answer.
There's a repeatable, rule-based actionprice calculation, first reply, lead routing, report assembly. Anything you can describe as "if X comes in — do Y" can be handed to a bot or a scenario.
There's somewhere to store dataa Google Sheet, a CRM (Pipedrive, HubSpot, Bitrix24) or at least email. Automation connects what you already have.

When it's still too early — honestly and upfront. If you get a handful of inquiries a month and each one is non-standard, it's cheaper to handle them by hand. We'll tell you that honestly on the call rather than sell you a scenario to tick a box.

Entry point: a simple integration — from €300. A custom AI agent — from €1,000. That's the cost of building it; you pay for third-party services separately (see the "Pricing" section).

Result

What we automate

Not a list of actions, but what ends up in your hands.

AI agent for handling inquiries

The bot receives an inquiry from your website or ads, qualifies the client, answers from your knowledge base and passes a ready lead into the CRM. A non-standard request goes to a live manager — the bot doesn't make things up. AI agent in the pricing →

AI quote and proposal calculation

From an inquiry the bot calculates the cost itself (materials, labor, options) and sends the client a ready proposal by email and on WhatsApp — in minutes, around the clock. You change prices yourself in a spreadsheet, no developer needed.

Chatbot on WhatsApp, Instagram and Facebook

Answers incoming messages at night and on weekends in the client's language, carries the conversation, escalates hot and gone-cold leads. Via TimelinesAI (WhatsApp) and ManyChat (IG/FB).

Lead delivery with nothing lost

An inquiry from any form → a card in Telegram with buttons + a row in a Google Sheet + an email to the owner. With source, date and contact. CRM and automation →

Make scenarios: connecting systems

Website ↔ CRM ↔ spreadsheets ↔ email ↔ messengers. Matching a quiz to the CRM by phone number, auto-translating content, callbacks based on lead quality.

Reports on autopilot

A weekly export from ad accounts into a spreadsheet and delivery to WhatsApp/email — without manual assembly.

Lead quality loops (Meta CAPI)

A manager marks "qualified / junk" in Telegram → the signal goes to Meta → the algorithm stops bringing in irrelevant inquiries. Works together with advertising. Analytics and tracking →

Reliability as part of the work

Retries on failures, a queue for lost inquiries, health checks, a kill-switch and Telegram alerts on outages. The raw lead is always duplicated to you — even if something goes down.

Pricing

Open prices, no "on request"

All prices are in EUR, excl. VAT. This is the cost of building and implementation; you pay for third-party services separately.

WhatPrice
AI agent for lead processingCustom agent: qualification, replies, handoff to CRM€1,000 – 3 000
AI workflow automationComplex Make or n8n flows with AI nodes€1,000 – 5 000
Simple integrationMake, Zapier or n8n: one process, 2–3 services.€300 – 800
Complex scenarioMake or n8n: multi-step logic, branching, error handling.€1,500 – 5,000

Why a range: it reflects the different amount of work within one service — how many channels we connect, how complex the bot's logic is, whether there's a knowledge base and how many systems we link. On the call we give an exact figure for your task.

  • How many channels we connect: just the website — or WhatsApp, Telegram, Instagram, Facebook.
  • Logic complexity: simple lead routing — or an AI agent with a knowledge base, quote calculation and qualification.
  • How many systems we link together: website ↔ CRM ↔ spreadsheets ↔ email ↔ messengers.

On third-party services — separately and transparently. Fees for WhatsApp transport, Make operations above your plan and AI tokens are separate from our work. At volume this is usually a few dozen euros a month.

The whole "AI marketing" section in the pricing → Integrations and CRM in the pricing → Take it all as a package with support →
Before → after

What changes after implementation

No charts or promises — what the routine looked like before automation and what's left of it.

Mzogs · fences, Latvia
A manager spent around 25 hours a month manually calculating quotes; inquiries at night and on weekends waited for an answer until morning.
The AI bot calculates the quote itself and sends a proposal in minutes, 24/7 — manual calculation dropped to almost zero.
Kronitek · B2B, Europe/US
B2B leads message on WhatsApp around the clock from all over the world — the team physically couldn't cover nights and weekends.
The AI sales assistant covers weeknights and full weekends, saving the team around 20 hours a week.
General principle
Inquiries got lost between the form, email and CRM; source and duplicates were sorted out by hand.
Every inquiry lands in Telegram, the spreadsheet and the CRM with source and contact — nothing gets lost.
Cases

Cases by niche

Not promises — real scenarios running in production.

MzogsFences · Latvia
Latvia · Make + Claude · Website, Facebook, WhatsApp

The manager spent time on manual quote calculations, and inquiries at night and on weekends waited for an answer until morning. We built an AI bot on Make and Claude: from an inquiry via the website or Facebook it calculates the full quote itself (panels, posts, gates, wicket, installation) and sends it to the client by email and on WhatsApp in minutes, around the clock. The owner changes prices himself — through a Google Sheet with an "Apply" button. The bot doesn't invent a price: for a non-standard order it hands off to a live manager.

25 → 0 hof manual quote calculation / mo
24/7a reply to every lead across 2 channels
selfowner changes prices in a spreadsheet
KronitekFood equipment · B2B
Europe, US, Canada, Australia · Make + Claude · WhatsApp, CRM

B2B leads message on WhatsApp around the clock from all over the world — the team physically can't cover nights and weekends. We built an AI sales assistant on Claude: during off-hours it replies in the client's language, qualifies (volume, product, geo), gives ballpark prices and timelines from the knowledge base, hands hot leads to managers by morning and writes a daily review of the conversations itself. The automated pipeline handles 150+ leads a week. The bot was fine-tuned on feedback from the sales team (20+ versions).

~20 h/wkof the team's time saved
150+leads a week through the pipeline
20+fine-tuning versions from feedback

On the call we'll show cases closer to your niche — with scenario logic and real numbers. When one scenario isn't enough and you need a separate system (like the booking system at Liberta) — that's already apps and systems.

Process

From task to a working scenario

A simple integration — 3–7 days. An AI sales rep with a knowledge base and fine-tuning — 3–6 weeks.

01
Day 1–2

We map the process

We look at where inquiries come from, where they get lost, what the manager does by hand. We capture the "if X comes in — do Y" rules. We show what's worth automating and what isn't yet.

02
Week 1

First working scenario

We build the most valuable piece — usually lead delivery with nothing lost, or quote calculation — and launch it on real data. You see the result, not a mockup.

03
Weeks 2–3

AI logic and channels

We connect the knowledge base and channels (WhatsApp, Telegram, IG/FB) and set up guardrails — rules under which the bot escalates to a human instead of inventing an answer.

04
Weeks 2–4

Fine-tuning on live conversations

The bot writes a daily review of its own chats with scores; we adjust the logic based on feedback from your sales team. We refine bots iteratively — usually 15–20+ versions until it works "like your best manager."

05
Onward

Support

Monitoring, alerts on failures, updating prices and the knowledge base, new scenarios as you grow. Control of prices and content stays with you — through a spreadsheet with a publish button.

Why us

Our commitments and rules

We don't just "set up bots" — we keep them in production for real clients and stand behind the fact that they don't go down.

Claude

AI on Anthropic's core

Calculation bots, conversational sales reps, spam filters, auto-translation — with "don't invent a price → escalate" guardrails. Not a cookie-cutter builder.

−54%

We save on operations

We brought one Make account down from ~14,000 to ~6,500 operations a month with no loss of functionality. You pay for results, not for bloated scenarios.

2 h

Engineering reliability

A reply within 2 hours (urgent — 30 minutes). Retries, a queue for lost inquiries, health checks, a kill-switch, Telegram alerts. Every quote is verified in tests.

Yours

Accounts and data

All accounts, access and scenarios are yours. You change prices and copy yourself in a spreadsheet. If you leave, you take it with you. No commission on your budget and no penalties for termination.

Questions

Frequently asked

Honest answers to what people ask before we start.

How is an AI agent different from a regular button-based chatbot?

A button bot leads people through a pre-drawn tree and breaks on any "off-script" question. An AI agent understands free text, answers from your knowledge base, calculates and qualifies. The key thing is that it has guardrails: if it doesn't know something or the request is non-standard, it doesn't make things up — it hands off to a human.

Won't the bot tell clients nonsense or quote the wrong price?

That's exactly why we set up guardrails. The bot takes prices from your spreadsheet, not "from thin air"; every calculation is verified in tests before launch. For a non-standard order the bot escalates to a manager instead of inventing a figure. Plus the raw lead is always duplicated to you — even if something goes wrong.

Will clients realize it's a bot?

That's your call, and we discuss it at the start. You can run the conversation transparently as an "AI sales assistant," or make the first reply seamless and hand anything complex straight to a human. We'll advise what's more honest and what converts better in your niche.

Can I change prices and copy myself, or do I have to come to you every time?

Yourself. Prices and content live in a Google Sheet with validation and a publish button — you change the numbers, hit "Apply," and the bot picks it up. No developer and no waiting.

And if a scenario breaks at night — will I know?

Yes. We build in health checks and Telegram alerts for API failures or outages. There's a kill-switch to stop the bot instantly, and a queue for lost inquiries that can be resent. Reliability is part of the work, not an option.

I already have a CRM / spreadsheets / a website. Will I have to change everything?

No. Automation connects what you already have: your website, CRM (Pipedrive, HubSpot, Bitrix24), spreadsheets, email, messengers. We plug into your infrastructure rather than forcing you to migrate.

How long does it take from start to launch?

A simple integration — 3–7 days. An AI agent with a knowledge base and fine-tuning on live conversations — 3–6 weeks: you see the first working scenario within the first week, then we tune the logic.

Is automation a one-off project or a subscription?

Building and implementation is a one-time payment within the range above. After that, support is optional: monitoring, updating prices and the knowledge base, new scenarios as you grow. The bot works without support too; support is needed when the system evolves and needs a hand.

We'll show what can be automated in your business

Tell us where inquiries come from and what the manager does by hand — in 30 minutes we'll map the process and show where automation pays off and where it's not worth it yet. Free and with no obligation.

Mzogs: 25 → 0 h of manual quotes a month · Kronitek: an AI sales assistant saves the team ~20 h/wk
We'll reply within 2 business hours. No spam. Straight to the point.