Marketing
5x content output, brand-consistent AI generation
- AI briefs, content planning
- Blog + social + email AI-assisted
- Style guide + example bank for consistency
- AI image generation in-house
AI workshops · 2026
AI isn't magic. We show you what it does — and what it doesn't.
Learn from people who ship RAG systems, AI agents and on-prem LLM deployments for a living. Not another ChatGPT tip list — repeatable workflows your team will actually run.
100 GB+ data processed · 80+ projects shipped · NIS2 + EU AI Act, daily
Why this is different
Builders teach. Not lecturers.
The market is full of "Master ChatGPT in 5 steps" courses. We do the opposite: we built a search engine over 100 GB+ of data for kultura.hu, run on-prem LLMs at NIS2-regulated clients, and deal with EU AI Act compliance daily. The workshops aren't about explaining tokens — they're about the time we lowered a quote because the AI agent would have been too expensive in production.
What AI can — and can't — do for your business
Most courses only cover column one. They skip column two. But useful AI rollout depends on knowing where it fails.
It can do this
It can't do this
Summarize structured data in 30 seconds
Guarantee numbers came from the source without RAG + verification
Draft emails 5x faster
Decide which offer to send a specific B2B customer
Answer questions inside a 200-page PDF (with RAG)200 pages
Provide legal advice — there's no liability holder
Categorize 10,000 support tickets in 2 hours10,000 tickets
Learn your internal jargon without fine-tuning or few-shot examples
Generate boilerplate code (CRUD, validation, reports)
Understand the business logic of your 8-year-old legacy system
Write 99% accurate English (95% Hungarian)99% / 95%
Match your brand voice without a style guide and example bank
Six modules. You use them the next morning.
Each module focuses on one workflow. We show how we do it. You do it. You go home and repeat it.
01Prompt engineering for non-engineers
How to give an LLM instructions so the 80%-useless answer becomes 80%-useful. Templates, structures, examples.
02AI for marketing teams
5x content in 1.5x time: blog, social, email, briefs, scripts. Brand-consistent generation with style guide and example bank.
03RAG knowledge base from your docs
How to assemble a chatbot over your own documents in a GDPR-safe way. Vector stores, chunking, hallucination control.
04AI agents in everyday work
n8n, Zapier, Claude Computer Use, Cursor — when each fits, when each fails. Real workflows your team can run.
05What NOT to do with AI
Real-world failures: hallucination, prompt injection, data leak, EU AI Act traps. Our own examples — including the ones where we got it wrong.
06Team-level AI rollout
How to roll AI into a 10–50 person team without chaos. Adoption planning, tool portfolio, measurement.
What your team takes home
Measurable difference within 30 days. Not a promise — straight from the playbook.
Content output
4 posts / week20 posts / week
Same marketing team, same quality bar.
Email response time
18 min / email4 min / email
Customer support and sales communication.
Document search
12 minutes30 seconds
RAG-based internal knowledge base.
New-hire onboarding
3 weeks9 days
AI assistant accelerates ramp-up.
Programs shaped to the team
Pick a focus or mix tracks — corporate package allows blending.
Sales
100x personalized outreach, deeper lead research
- AI assistant inside CRM
- Lead research automated
- Personalized email sequences
- Demo prep with AI
Operations
Document processing, email triage, report generation
- Inbound document processing
- Email categorization and triage
- Report generation from spreadsheets
- SOP digitization
Engineering
Cursor + Claude Code at master level, AI in code
- Advanced Cursor / Claude Code workflow
- AI agent in CI/CD
- Code review automation
- RAG over your own codebase
One day, six modules, tangible output
Scroll through — this is the spine of the 1-day workshop. The corporate package is modular and can run multi-day.
- 01
Opening — what AI can and can't do in 2026
A realistic picture of today's models. What they're good for, what they aren't. Real production examples from client projects.
- Model families (Claude, GPT, Gemini, Llama)
- Cloud vs on-prem
- Cost model and turnaround
- 02
Prompt engineering, hands-on
Working on your laptop. We give you a concrete task from your domain, and iterate the prompt until the result is good. Structured input, examples, output format.
- Templates and anti-patterns
- Few-shot examples
- Output validation step
- 03
Build your own workflow
Pick a workflow from your daily work. We build it together — template, tool, validation. You use it the next morning.
- Selection: what's worth automating
- Tool choice (n8n, Zapier, custom)
- Error handling and testing
- 04
Compliance & security
GDPR, EU AI Act, NIS2. What you can't do in cloud. How to do it right. Real examples from clients and from our own missteps.
- Data classes and AI tools
- Prompt injection and data leak
- Logging, audit, explainability
- 05
Team rollout
Who, when, in what order. Cost, measurement, success criteria. A 30-day plan you take home.
- Adoption strategy
- In-house tool portfolio
- ROI tracking and reporting
- 06
Q&A + 30-day support starts
30 minutes dedicated Q&A. You get the playbook. Slack channel for 30 days, plus a follow-up Q&A 2 weeks later.
- Playbook (PDF + Notion)
- Slack support, 30 days
- Follow-up session in 14 days
How we read a contract
An AI agent finds the critical fields in 12 seconds. You watch us do it, then run the same on your own document.
Supply contract #2026/0412
AI agent · 12s- Buyer: Example Ltd. (VAT: 12345678-2-41)→ VAT
- Supplier: AppForge Solution Kft.→ Vendor
- Contract value: 1,250,000 HUF + VAT→ Net value
- Payment due: 2026-05-15→ Due date
- Late fee: 0.5% / day, max 30 days→ Penalty
- Jurisdiction: Budapest Metropolitan Court→ Disputes
6 fields in 12 seconds. At the workshop you run it on your own document.
Three ways to run it
Individual, team, or whole organization — same experience behind all three.
1 day (4-hour intensive)
Public workshop
Individual signup, max 12 seats
€129 / person
- Live hands-on tasks
- Playbook + 30-day Slack
- Follow-up Q&A in 2 weeks
- Certificate
1–2 days, in-houseNépszerű
Corporate package
6–20 person team
from €2,200
- Modules tailored to the team
- On-site or online
- Team-level playbook
- 60-day support, 2 follow-ups
4–12 weeks
Custom program
Whole-org integration
Custom quote
- Workshop + coaching + rollout
- AI tool portfolio rollout
- Measurement and audit framework
- Dedicated advisor
Frequently asked
Yes. 4 of the 6 modules are fully non-technical. "Prompt engineering" and "AI for marketing teams" are specifically designed for non-technical teams — typical attendees are a marketing manager or a managing director.
Ready?
AI is only worth something if your team knows how to use it.
One day. We show your team what AI does — and what it doesn't. Let's get the details right.