Mykhailo Lohaza

I automate routine work in a way people actually want to use.

I believe that in the age of AI, people matter more than ever.

For five years I've been rolling out CRM and ERP systems, and since 2024 AI tools too, and I keep seeing the same thing. The system works, but people find ways around it and drift back to Excel. The tool is rarely the problem. Usually nobody asked the people who'd be using it.

On an ERP rollout at a 300-person company, we migrated 600–800k rows of HR and finance data without critical loss and finished 1.5–2 months ahead of schedule. What made it work, though, was 9–10 training sessions and key people in every department. In the end, 90% of staff started using the system.

I bring the same approach to AI. During 2.5 years in the pharmaceutical industry, I launched an AI agent for Jira as Product Owner. Managers brief it in plain language, and 10–15 people used it in the pilot. Code handles the simple commands and AI only does the analysis, which cut AI costs by 87%. The AI never made up data when writing to Jira. I build automations in n8n and Make, AI agents and assistants that answer questions from your own documents, and I help teams set up Claude.

Today I help agencies, companies and sole traders automate routine work without layoffs, using my own Calm Ops method. I listen to the team first, automate what they themselves call routine, train people and report the results in hours saved.

Since 2020

rolling out CRM and ERP; building AI agents since 2024

Poznań

working remotely with teams across Poland and the EU

Three languages

English, Polish, Ukrainian

Book a call30 minutes, €45

Five projects I'm proud of

I don't name clients, so each case shows where things started, what we did and what changed.

AI agent for Jira at a pharma SaaS company. My role: Product Owner.

Jira you can talk to

Before: Managers updated tasks and risk reports in Jira by hand.
What we did: Built an AI agent you brief in plain English. Code handles the simple commands, and AI does the analysis.
Result: AI costs down 87%, and the AI never made up data when writing to Jira. 10–15 people in the pilot.

ERP at a 300-person HR tech company. My role: rollout and training.

A new system people actually use

Before: Pulling the board report together took 4–5 hours, and HR and finance data sat in legacy systems.
What we did: Migrated 600–800k rows of data without critical loss, ran 9–10 training sessions and found key people in every department.
Result: The report now takes 5–10 minutes, 90% of staff use the system, and the project finished 1.5–2 months early.

HR agency with 5 departments. My role: CRM rollout.

From spreadsheets to one CRM

Before: Every department kept its own candidates and clients in separate Excel files.
What we did: For the first month I listened to the teams and noted what they called routine. Then we moved every department onto one CRM.
Result: None of the five departments went back to Excel.

Content factory for a detective game brand.

Stories nobody makes by hand

Before: Every Instagram Story was made by hand.
What we did: A chatbot collects the brief, and AI writes the content plan and generates the backgrounds to set parameters.
Result: One brief turns into 30 ready-to-post Stories, and the system keeps producing new batches.

Complete.dev hackathon, February 2026. Team of two, 48 hours. My role: product, architecture and the first agent.

A team training scenario in 90 seconds instead of 5–10 hours

Before: A trainer spends 5–10 hours writing a team training scenario by hand, and nobody processes the participants' questionnaires.
What we did: In 48 hours the two of us built LARP Orchestrator. The team fills in a questionnaire, one AI agent finds the weak spots using Lencioni's five dysfunctions model, and a second one writes a role-play game: 4–9 scenes, characters and a guide for the facilitator.
Result: A ready scenario in 90 seconds, and 2nd place for Best Business Solution among 70+ teams and 300 participants from around the world.

More case studies on LinkedIn I write about rollouts, automation and how AI is changing the way teams work. View profile

Experience

2020–2021

CRM for an HR agency: five departments moved off Excel onto one system.

2022–2024

ERP at a 300-person HR tech company: data migration and team training.

Since 2024

Product Owner of an AI agent for Jira at a pharma SaaS company. 2.5 years in pharma overall.

Hackathons

Complete.dev: 2nd place among 70+ teams with LARP Orchestrator. Also the Software Mansion × Gemini hackathon.

Manual work to agents, decisions to people.

Straight talk about tools, real care for people.

Three rules

A Human Gate at every step. Agents move data, write drafts and send reminders. Anything that touches money or people gets approved by a person. I call that checkpoint the Human Gate.
Listen to the team first. I automate what people themselves call routine, which is why they actually use the automations afterwards.
Report in hours saved. After the pilot you see how many hours came back and how many people really use the automation.

What I don't do

Projects aimed at cutting jobs.
Promises of results without a pilot, or automations nobody will maintain.

Education

PMI-ACP
MSc Business Intelligence & Big Data
MSc Management, Adam Mickiewicz University (UAM)

Not sure where to start?

If AI still feels more like a worry than a help, book a call. In 30 minutes we'll work out where your team should begin.

Pick a time

€45

for 30 minutes