AI business process automation: routine work that runs on its own
I automate the repetitive work in your company with AI: incoming requests, documents, reports, follow-ups. We start with the process that costs the most time, I set it up on the tools you already use, and the people who run it learn to look after it. You pay by the hour, and the first call is free.
Signs a process is ready to automate
- Someone copies the same information from emails or forms into a CRM or a spreadsheet every day.
- Documents arrive as attachments and are read, sorted and retyped by hand.
- A weekly report takes half a day of collecting numbers from different places.
- Customers ask the same questions, and the answers are already in your documents.
- Follow-ups and reminders happen only when someone remembers them.
What I automate
- Incoming requests and emails: read, sorted, routed to the right person, with a draft reply ready to check.
- Documents: invoices, contracts and applications read by AI, with the data filed where it belongs.
- Reports: assembled on a schedule from your CRM, accounting and spreadsheets, and sent to whoever needs them.
- Follow-ups and reminders: to customers, suppliers and your own team, on time and in your tone.
- Customer questions answered from your own documents, with a hand-over to a person when needed.
- Data moving between your systems without anyone copying and pasting it.
AI where reading is needed, rules everywhere else
A good automation is mostly ordinary steps: when this happens, do that. AI is the step that reads and writes — an email in free text, a scanned invoice, a question phrased a hundred different ways. I use it only for those steps, and wherever a mistake would cost money or trust, a person confirms its work before anything goes out.
The tools
Workflows in n8n or Make; Claude or ChatGPT through their APIs for the reading and writing; and the CRM, mailbox, spreadsheets and messengers you already use. Where data must not leave the company, the AI runs on your own servers. Everything is set up in your company's accounts.
What I do
Map the process
With the person who runs it, step by step: what comes in, what is decided, what goes out, and where time is lost.
Build the first automation
On your tools and accounts, starting with the steps that save the most time at the lowest risk.
Run it alongside the old way
For a while both run, so we can compare results and fix what the real data reveals.
Hand it over
I document what runs where and teach the person who owns the process to adjust it and to notice when something goes wrong.
Take the next process
Once the first one is part of the daily work, we choose the next by the same measure: time saved against risk.
Who you work with
Sergiy Yashchenko
AI consultant · 20+ years in IT · Kyiv
I was deputy director of the AI department at a telecommunications company with 3,000+ employees, where automation had to work at scale and within strict data rules.
As a CEO, I automated the work of my own companies — choosing what to automate first and setting much of it up myself.
An example
Related services
Price
$50 per hour — the same rate as everything else I do. The first 30-minute call is free: on it I estimate how many hours your task takes, and you can stop after any hour.
Frequently asked questions
Which process should we automate first?
The one that is frequent, takes real time and is low-risk if something goes wrong. We usually find it on the free call, from a short description of your week.
Do we need new software?
Usually not. The automation connects what you already use. You may need a workflow tool such as n8n or Make and an AI subscription, both in your company's name, and I tell you what they cost before we start.
What if the AI makes a mistake?
Automations are built for it: where a mistake matters, a person approves the AI's work before it is sent, and every run is logged so it can be checked and corrected.
Who looks after the automation afterwards?
The person who owns the process. I document it and teach them to adjust it. When you need something bigger, you book another hour.
Is our data safe?
We choose the setup by how sensitive the data is: business plans and APIs that do not train on your data, or models on your own servers. Your IT keeps control of access.
How much does it cost?
$50 per hour. The tools are paid separately, in your company's name. The first 30-minute call is free.