Prezentare
20–30 minutes
A topic, examples from live systems, and the numbers we can show. No product slides.
Company
We come with what we have built, not with a pitch. An AI voice agent someone from the audience can talk to, the figures we can show and the lessons it cost us. Tell us the audience, format and date.
Formate
Choose the option that fits your schedule. If none fits, still write to us with what you need.
20–30 minutes
A topic, examples from live systems, and the numbers we can show. No product slides.
10–15 minutes
A voice agent that someone in the room talks to, live. It needs internet access and sound in the room — see the conditions below.
60–90 minutes
Participants leave with something built: an agent that responds on their data or a connected workflow.
as required by the agenda
A discussion with other guests. We prepare on the topic, not only on what we sell.
one day or a weekend
Teams building with agents: architecture, model choice, what breaks at the first real user.
to be agreed
For teams evaluating AI before investing. Format and conditions, agreed in writing beforehand.
Teme
Each topic comes from something that is running or has run with us. If you want a topic that is not here, ask — we will say frankly whether we have anything to show or not.
What sub-second latency means, why the conversation breaks off, where transcription still makes mistakes in our languages, and what we changed after each failure.
Systems that work in a demo and fail in production: overly talkative agents, unsynchronised catalogue, down phone lines. The lessons cost more than the successes.
How we divide the work between people and agents in our own company, what we automate and what we refuse to automate.
Copernicus observations, resolution, revisit frequency, and why a surface model is not a terrain model. The data are public and European — we build the layer on top.
What the new data protection law in the Republic of Moldova means for a system that talks to clients: legal basis, transfers, retention, the person's rights.
How to choose the problem, how to prepare the data, how to measure whether it worked, and when it is right to stop the project.
Istoric
Only the events where we took the stage, had our own stand, or entered the competition. Attendance from the hall is on the history page.
Lilia Chetrean, speaker; Oleg Chetrean, CEO & Founder Mega Promoting
Lilia Chetrean, for aichat.md
Oleg and Lilia Chetrean, under the name Agintify
Oleg and Lilia Chetrean, for AIChat / Bravin
Lilia Chetrean
Kallina AI
Lilia Chetrean, speaker; Oleg Chetrean, SME representative
Oleg Chetrean
Conditions
Nothing unusual. Just a few things agreed in advance, so the demonstration does not fall flat in front of the room.
Travel, duration and conditions are agreed together, in writing, before confirming. We do not sell from the stage: if someone in the room wants to work with us, we talk afterwards.
The date, the audience, the format and the language. We reply with our availability and with what exactly we would bring into the room. Or write straight to hello@megapromoting.com.