Mykola & Egor Discuss a task

AI implementation for business processes · audit · automation

AI that pays off. The rest is not our job.

We are Mykola and Egor. We come into your company, take processes apart operation by operation, and bring in AI only where it saves hours and makes money. Payback is calculated before we start, and we answer for it with numbers in your CRM.

Get a process review or message us on Telegram ↗

We take on at most two new projects a month. The two of us do the work, not a team of interns.

report · process diagnostics No. 11-D
  • Incoming order processing 4 h → 12 min
  • Payment reconciliation manual, 9 h/wk
  • Quotes and proposals 2 days → 40 min
  • A chatbot “like everyone has” won’t pay off
  • Package payback 2.5 mo

total per map: −1,100 hours of manual work a year

Over two years: 11 implementations. Median payback is 4.2 months. Every euro invested returned ×3.4 in the first year.

01 Diagnosis p. 02 / 07

Most “AI adoptions” never pay for themselves.

A chatbot nobody opens. A neural network “for the image”. Subscriptions to a dozen tools that three people use. We have seen it from the inside, which is why we work in a different order: count the money first, write the code second.

Where the money usually leaks:

  1. 01

    Orders wait hours for a reply, and the client goes to whoever answered first.

  2. 02

    Managers spend half the day moving data between Excel, the CRM and email.

  3. 03

    Quotes take days to calculate, and deals go cold before the numbers arrive.

  4. 04

    Reports are compiled by hand at month-end, when it is too late to change anything.

If you recognized your business in at least one item, we most likely have a case with the math on exactly that. Three of them are below.

02 Case studies p. 03 / 07

Three projects whose math we are not ashamed to show.

CASE-01 Building materials distribution · 120 employees

Processing one incoming order took:

4 hours 12 minutes

AI reads incoming email and website orders: it recognizes line items, checks stock and fills the order in the ERP. The manager only confirms. The routine part of the job is gone.

I didn’t believe the numbers in the report at first. We recalculated ourselves, and they checked out.

commercial director, name on request
Managers tied up in processing
3 → 0.5
Lost orders
~7% → 0.4%
Manual work per year
−1,100 h

paid off in 2.5 months ✓

based on the client’s ERP and CRM data over the first three months after launch

CASE-02 Retail chain · 14 stores

Cash frozen in inventory:

−23% in six months

Demand forecasting per store and automatic ordering from suppliers. Purchasing used to be planned on the chain director’s gut feeling, and the money sat on shelves as dead weight.

Purchasing forecast accuracy
60% → 87%
Expired stock write-offs
−31%
Freed from inventory
≈€118,000

paid off in 4 months ✓

based on the client’s inventory system reports over six months

CASE-03 Furniture manufacturing · 80 employees

Preparing a quote used to take:

2 days 40 minutes

AI calculates the estimate from drawings and specs, then assembles a proposal using the company’s template. The client gets numbers first, before any other contractor, and tends to pick whoever answered earlier.

Quote-to-deal conversion
+19%
Estimator workload
100% → 30%
Quotes per month, no hiring
×4

paid off in 3 months ✓

calculated together with the client’s CFO from the margin of closed deals

03 How we work p. 04 / 07

Pilot first. The risk is on us.

Every stage ends with a document that stays with you, even if we don’t work together afterwards.

  1. 01

    Diagnostics · 1–2 weeks

    We sit inside your processes: where people burn hours, where orders get lost, where mistakes cost money. The output is a process map with the math: what to automate, what to leave alone, and how much it will bring.

  2. 02

    Pilot · 2–4 weeks

    We take one process and drive it to a result for a fixed price. Success criteria are written down in numbers before we start. If the numbers don’t add up, we stop, and the map stays with you.

  3. 03

    Rollout · 1–3 months

    We scale what proved its payback in the pilot: integrations with your CRM, ERP and email, team training, working procedures.

  4. 04

    Support · on request

    We watch the metrics, retrain the models, and send a monthly report with the same numbers you used to make the decision.

04 Who we are p. 05 / 07

There are two of us, and that is deliberate.

You work with us directly. No account managers, no subcontractors: the people on this page are the same people who will sit inside your processes and write the code.

Mykola

Mykola

processes and models

“Ten years in data, from retail analytics to ML in production. My part is finding the processes in your operational chaos that can be counted, and building a model that won’t fall apart in a month.”

Egor

Egor

integrations and rollout

“My job is making sure AI survives to production: integrations with the ERP, CRM and email, team training, and metrics that show it actually works.”

05 Questions p. 06 / 07

Honest answers to awkward questions.

How much does it cost?

Diagnostics is a flat €2,500 and takes up to two weeks. A pilot starts at €6,000, with the price fixed before we begin. Rollout is priced from the process map: after diagnostics you know both the cost and the payback forecast, and only then decide whether to continue.

Our data can’t leave the building.

A normal situation. We deploy open models inside your perimeter, so the data never leaves your server. Roughly half of our projects were built this way.

What if it doesn’t work?

Before the pilot starts we fix the success criteria in numbers. If they aren’t met, we stop and you don’t pay for the rollout. The process map and all the calculations stay with you.

We’re not an IT company, we run a “normal” business.

Even better. The fastest payback hides in “boring” processes: orders, purchasing, quotes, reconciliations, reports. Wherever people spend years doing by hand what a machine does in seconds.

Who are you a bad fit for?

If you need “AI for a board presentation”, that’s not us. And if the company has fewer than ~15 people, the economics of automation usually don’t add up. It’s more honest to say so upfront than to take the money.

06 Contact p. 07 / 07

Tell us where it hurts.

The first call is 30 minutes, free and with no strings attached. We’ll look at your processes and tell you honestly whether there is money here. We reply within one business day.

Fastest: Telegram @your_username · message us, straight to business