Private workshop
Testing AI before you trust it.
A one-day workshop for teams that build, buy or approve AI tools. Your team leaves with a test set built from its own questions, clear pass levels and a release checklist it can use the next morning.
$3,500 or €3,000 per team of up to ten. One day, on site or online.
Book a 30-minute callWho it is for
Operations, quality, engineering, IT and product teams who:
- are piloting an AI assistant or agent and need to know whether it is ready
- are choosing between AI vendors and want a fair way to compare them
- approve AI tools for use and want a clear basis for that decision
What your team leaves with
A test set from your own work.
A first set of 50 real questions your people ask, each with the approved answer and its source.
Pass levels that match the risk.
Stricter levels for safety, quality and customer-facing answers. Clear levels for everything else.
A scoring sheet.
A simple way to score any AI tool against your test set, answer by answer.
A one-page release checklist.
What must be true before an AI tool goes to more users, and who signs it off.
A record of the training.
An attendance record with the topics covered, for your internal AI training records.
The day
Morning 1: How AI tools fail.
Wrong answers stated with confidence, missing sources, outdated documents, and answers to questions the tool should decline. Live examples on public manuals.
Morning 2: Building your test set.
Your team turns its own questions into test cases with approved answers and sources.
Afternoon 1: Scoring and pass levels.
Score a real AI tool against the test set. Set pass levels by risk.
Afternoon 2: Release decisions.
Build the release checklist, assign who approves what, and plan retesting after changes.
Six working hours with breaks. Up to ten participants.
Where this method comes from
This is the method behind an assistant built inside a Fortune 500 manufacturer. It turned 4,000+ technical documents into answers engineers could use in seconds, reaching 97% accuracy across seven product lines and 100% on critical equipment, tested on about 2,000 real questions.
The workshop is led by Virginia Kalesić, founder of AI Tech Magic. She built and led a global team of 15+ AI developers, data scientists and AI engineers, set up AI governance and a stage-gated path to production, and ran an innovation lab that cut AI development cycles from 18 to 24 months to 2 to 4 weeks.
What to prepare
- 20 questions your people ask in their daily work
- The AI tool you use or are considering, if you have one
- One person who can confirm the correct answers to your questions
Please share no confidential documents before an NDA is signed.
Price and terms
- Price
- $3,500 or €3,000 per team, fixed
- Group
- Up to ten participants
- Format
- One day, on site or private online session
- Payment
- 50% at booking, the balance after the workshop day
- Travel
- On-site delivery outside Tuzla adds travel at cost, agreed in advance
- Next step
- Teams that start an Assistant Pilot within 30 days receive €1,000 off the Pilot
Questions
Do we need technical staff?
No. The workshop works for mixed teams. Engineers and IT staff go deeper in the scoring session.
Do we need an AI tool already?
No. Teams without one use a public example and leave ready to test any tool they consider.
Is the session recorded?
Only if you ask us to. Online sessions are private to your team.
Can it be delivered in Bosnian?
The workshop is delivered in English. Bosnian support can be arranged for local teams.
Give your team a fair test for any AI tool.
Thirty minutes is enough to agree the date, the group and the questions to bring.
Book a 30-minute call