We bought licenses. Nobody uses them.
The company invests in AI tools, then a month later everyone is back in Excel. Lack of adoption isn't a tech problem — it's a people problem.
// 01 — Mission
Most AI rollouts end the same way: licenses sit unused, Excel comes back, the deck ends up in a drawer.
I'll show you how to do it differently.
Trusted by:1200+ trained55+ organizationsAmazon · Accor · 4 countries
Scroll
// Manifesto
AI won't changeyour company.Peoplewho knowwhat to do with it— will.
— Przemek.
After twelve years in the industry, after two years with generative AI.Challenges
I know.
I spotted the pattern two years ago. Started keeping a notebook. I call it "Sentences I've heard before."
Six of them come up so often I already know what I'll hear next time.
The team is scared AI will take their jobs.
Resistance is natural. But without working through it, every rollout ends up as slides in a drawer. People need to see the benefit, not the threat.
We don't know where to start.
ChatGPT, Copilot, automation, agents — too many options, not enough context. Without a process diagnosis you end up picking a tool for a problem you don't have.
Training happened. Results didn't.
A one-off workshop with a generic deck doesn't change habits. Real change needs a program matched to your processes, not to the vendor's slides.
Leadership wants results but doesn't get AI.
Decisions get made at the top, knowledge sits at the bottom. Without a shared language between leadership and the team, the rollout has no sponsor — and dies within a quarter.
Information overload.
The AI landscape moves so fast that by the time IT approves a tool and the business reacts to the change, the market has already moved on.
I'll say something not everyone wants to hear: technology has nothing to do with it.
Each of these six problems is a story about people, habits and fear. That's what I work on.
// 03 — Who I am
Przemek
Gliwice · PL·Since 2013
Trainer · Strategist · Psychologist
Przemek
Blicharski.
I started in sales and analytics, back when "AI" was still an academic conference topic, not a boardroom one. Twelve years in data & AI taught me one thing: technology without understanding people is just an expensive toy.
That's why I went back to study psychology at SWPS — not to change careers, but to understand resistance, motivation and how teams actually take on new tools. The combination of hard data and the soft side of change defines everything I do.
I rolled out BI and data strategies across Europe and the US — Qlik, Power BI, Snowflake, SAP. When generative AI changed the game, I knew exactly where the bridge between technology and business was missing. I stepped onto that bridge.
Today I run PBX Agency, deliver AI training and implementations for global organizations including Amazon and Accor, lecture at SWPS and SGGW, and co-organize AI Breakfasts Silesia. 1200+ people trained across 55+ organizations in four countries. I don't sell AI — I teach how to use it so it actually works.
“AI won't take our jobs.
AI will take the drudgery of work.”
I studied at
IBM·Google·Johns Hopkins·Penn·UCI·University of London·TU/e Eindhoven·CFI
// 04 — What I doThree ways
Three ways
to work together.
For teams and managers
AI Training
Hands-on 4–8 hour workshops delivered for teams at Amazon, Accor, Maspex and 50+ more. Your team walks out with skills, not a slide deck. 70% of the time is exercises on real tasks.
- AI fundamentals & prompt engineering — from zero to independent
- Advanced use: automation, data analysis, content creation
- A dedicated program matched to your industry and team's level
- Take-home materials + 2 weeks of email follow-up
For executives and decision-makers
AI Implementation
The full path: audit → pilot → team independence. I don't roll out tools — I roll out a change people accept. Resistance gets worked through, not ignored.
- Process audit and 3–5 high-impact AI opportunities identified
- Implementation map with budget, timeline and KPIs
- Pilot on a single process — measurable results in 2–4 weeks
- 3 months of post-launch support included
For event organizers
Keynote / Talk
I don't read slides. I provoke thinking, share real case studies, and leave the audience with questions that keep working weeks after the event.
- Topics: AI implementation, future of work, data-driven culture
- Formats: 30–60 min keynote, panel, conference workshop
- Track record: Data Science Summit, ICAN Institute, Amazon, Accor
- Delivered in Polish or English — EU, Nordics, US markets
Six steps
to results.
No shortcuts. But there is a road without the bullshit. I show you what I do at every stage — so you know what you're paying for.
Consultation
We start by analyzing needs — people, processes, goals. This is about understanding your context, not a sales call.
Diagnosis
I identify 3–5 areas where AI delivers measurable impact. Focus on real needs, ruling out everything that wouldn't fit.
Strategy
You get a clear plan: what, who, when and at what cost. The roadmap is yours to keep — whether or not we work together.
Implementation
I run the training, configure the systems, launch the pilot project. Your team gets the skills they need to own it.
Support
For the next 3 months I monitor results. We optimize processes, fix problems, remove blockers.
Independence
I build skills inside your company. The goal is a team that runs AI solutions on its own — without permanent outside help.
// 06 — ResultsNumbers
Numbers
that speak.
0+Training participants
0+Organizations
0Countries
Przemek has a gift for explaining complex things in plain language. After the workshop my team actually started using AI — not just talking about it.
Project managerManufacturing company, Silesia
I didn't expect an AI training could be this practical. Zero theory for theory's sake. Just tools I now use at work every day.
Marketing specialistDigital agency
Engagement
OrganizerAI Breakfasts Silesia
Recurring meetups for AI practitioners in the Silesia region
Speaker & Program CommitteeData Science Summit
One of Central Europe's leading data science conferences
Lecturer, AI programsSWPS & SGGW
Teaching AI in postgraduate programs
Trusted by
Amazon
Maspex
Accor
PZL Mielec · Lockheed Martin
FM Logistic
Polpharma
SGGW
Data Science Summit
Amazon
Maspex
Accor
PZL Mielec · Lockheed Martin
FM Logistic
Polpharma
SGGW
Data Science Summit
Altkom Akademia
Compensa
Budimex
Nationale-Nederlanden
Vector Consulting
WCK Wałcz
OTReK Ostrołęka
ICAN Institute
Altkom Akademia
Compensa
Budimex
Nationale-Nederlanden
Vector Consulting
WCK Wałcz
OTReK Ostrołęka
ICAN Institute
// 07 — ThoughtsI write what I think.
All articles →I write what I think.
Not what's expected.
First posts coming soon.
Frequently asked
questions.
How much does an AI training cost?
Pricing is set individually after a free diagnostic call — no two AI programs are the same. For international clients, quotes are provided in EUR or USD on request. The field moves fast: I update my programs every week, spend dozens of hours monthly tracking new tools and testing new models. What I taught three months ago looks different today — and that's the point. Your team gets what actually works now, not last quarter's playbook. Invoiced with VAT, 14-day payment terms. Preferential rates for NGOs and startups.
Can AI training be subsidized or funded externally?
Yes — depending on your country and company size. In Poland, SMEs can apply for up to 80–100% subsidy through government training funds (BUR, KFS) or EU regional programs. For companies elsewhere in the EU, similar co-financing schemes often exist at national or regional level. I refer Polish clients to a trusted partner who handles all the paperwork. For international organizations, training is typically invoiced directly — ask during the first call and we'll work out the best setup for your situation.
Where should we start with AI implementation?
From a process diagnosis, not from picking a tool. The most common mistake: companies buy ChatGPT Enterprise or Microsoft Copilot licenses before they've diagnosed where AI actually saves time. My process: (1) a two-hour strategic workshop with leadership, (2) identification of 3–5 highest-ROI processes, (3) risk & compliance mapping (GDPR, EU AI Act), (4) tool selection for the diagnosed processes — only now. Full methodology described in the "How I work" section.
Do you run AI training on-site or online?
Both formats. On-site anywhere in Europe and beyond — I travel for workshops at corporate headquarters, conference venues and client sites. Online for distributed and international teams (in Polish or English). I prefer on-site — group dynamics, team exercises and networking effects are much stronger that way. The online format works well for hybrid organizations and shorter follow-up modules.
Who is this for? Do I need to know AI?
You need to know nothing. Most participants of my workshops had never consciously used ChatGPT or Copilot before the training. Every program is matched to the group's level and industry — a B2B sales workshop looks different from one for manufacturing execs, IT teams or HR. I also have advanced programs for companies already using AI who want to go deeper: agents, n8n/Make automations, Claude and ChatGPT workflows.
What sets you apart from other AI trainers?
Three things. (1) Twelve years in data & analytics before AI became fashionable — I rolled out BI, Qlik, Power BI, Snowflake across Europe and the US and delivered AI programs for global organizations including Amazon and Accor. I know what technology adoption in a large organization actually looks like. (2) I'm a psychologist by training (SWPS University) — so I understand team resistance, adoption mechanics and why 80% of IT projects end up in a drawer. (3) I don't sell hype. I show tools that work on Monday morning.
Which AI tools do you teach?
Foundation: ChatGPT (OpenAI), Claude (Anthropic), Microsoft Copilot, Google Gemini, Perplexity. Specialist tools selected by context: Midjourney and DALL-E for imagery, Suno for audio, NotebookLM for working with documents, n8n and Make.com for process automation, Cursor and Claude Code for technical teams, Agent SDK for agentic deployments. I don't lock into any single vendor — I match tools to your processes, not the other way around. If you already pay for Microsoft 365 or Google Workspace, we'll use what you already have.
How long does an AI implementation take?
First visible results: 2–4 weeks from start. Full implementation with 3 months of support: a quarter. I don't run year-long projects — in AI, six months is a different world (new models, new capabilities, new pricing). We work in short iterations on specific processes — sales, customer service, reporting, HR, marketing, operations. After each month we review what works, what to change, where to scale.
Won't the team be afraid AI will take their jobs?
Often, yes — and it's good when that question gets raised. I don't fight resistance, I work with it as a psychologist would: a conversation is the first part of the workshop, before we even open any tool. Experience with 1,200+ trained people in 55+ companies: after a month of regular AI use, the strong majority say AI took the drudgery out of their work, not the work itself. There's always a group of skeptics — that's fine too; their arguments often catch real risks the enthusiasts miss.
How does the engagement work formally — contract, VAT, NDA?
PBX Agency (based in Gliwice, Silesia): VAT invoice, 14-day payment, milestone payments possible on larger projects. NDA is standard before the diagnostic call if the project needs one — I have my own template or I'll sign yours. Process: free 30-min call (Google Meet or phone) → diagnostic workshop → written proposal and quote → delivery → 30-day follow-up. No paperwork theatre, no lawyers cc'd, no five-round negotiations. Direct person-to-person.
// 09 — ContactWrite.
Write.
You don't
need to be ready.
Most companies I work with started with one email asking "I'm not sure this is even for us." That's the perfect place to start.
Free diagnostic call. I reply within 24 hours.