Frappua

Frappua Venture studio building practical AI, XR and digital products. Kahviauto, Liikkuva kahvila

23/10/2026

One teacher, thirty students, forty-five minutes. There is no time to wait for each learner or to go to each one, so thirty students adapt to one teacher. Not because of the teacher, because of the numbers. This is slide 4 of the deck where we learn in public and ask what changes when VR and an AI mentor give every learner their own time. The voice is synthetic, reading our text. All 12 slides are on our site.
https://frappua.win/safeskillvr/pedagogy

22/10/2026

How a group answers one customer in Next, please! On the right wall the team picks an answer, and it takes two touches on the same card to choose it, so the group has to agree. Then one student, the Speaker, says it aloud in their own words, face to face with the customer, and taps Said.
https://frappua.win

Almost all educational AI is built to answer faster. In a study with almost 1,000 high school students and an AI maths t...
21/10/2026

Almost all educational AI is built to answer faster. In a study with almost 1,000 high school students and an AI maths tutor, students solved more while they had the tutor, and without it they did worse than students who never had one. A tutor that gave hints instead of answers avoided most of the harm (Bastani et al., 2025, PNAS). In SafeSkillVR we want the opposite direction: the student tries, fails and searches, and the mentor stays close without taking the question away. This is how we design the mentor, not a measured result.

20/10/2026

In Next, please! the room turns into a car-service reception and a life-size customer walks up to the group. Anna is in a hurry. The left wall shows her mood in colour, shape and word, what she needs, and a body-language cue: she checks her watch. Students practise reading a customer before they meet a real one.
https://frappua.win

What does SafeSkillVR actually adapt to the learner? In our working teaching model we list six things: the order of expl...
19/10/2026

What does SafeSkillVR actually adapt to the learner? In our working teaching model we list six things: the order of explanation, practice, feedback and reflection; how much help to give and what kind; how much support to keep and when to remove it; when the mentor steps in and when it stays quiet; how much independence the trainee gets; and when to repeat, make it harder or move on. None of this works without a model of the learner. Without it, adaptive training is only a branching script. We have not yet marked which of the six already run in the prototype. That is the next team discussion.
https://frappua.win/safeskillvr

18/10/2026

Next, please! is a group game for an immersive room. The room is four projected walls, about 4.5 by 2.6 metres each, with one 360-degree scene, multi-touch on every wall and 5.1 surround sound. Nobody sits at a screen: the whole group stands inside the scene together. Shown here as a visualisation of the room.
https://frappua.win

How we want the mentor in SafeSkillVR to help. First the student tries. If that is not enough, the environment gives a s...
17/10/2026

How we want the mentor in SafeSkillVR to help. First the student tries. If that is not enough, the environment gives a signal: a sound, the needle, steam. Then the mentor's body does it: looks there, walks over. Only after that comes a question, then a hint in words, and the answer at the very end. Support first, then step back. These are design intentions, not measured results. The full slides are on our site under Learning in public.
Link in bio: frappua.win

07/09/2026

SafeSkillVR: VR training built around your procedure, not a catalogue course. Knowledge check, skills training, safety briefing, sustainable practice. The mentor guiding the trainee in the video is an AI inside the simulation. Built by Frappua! Link in bio: frappua.win

07/09/2026

We build VR training simulations around a company's own procedure. This is SafeSkillVR: knowledge check, skills training, safety briefing, sustainable practice. The woman in white is an AI mentor living inside the simulation. frappua.win

31/08/2026

Episode 01 of what we build at Frappua!
We design and build video monitoring systems where cameras understand what
they see: you describe a scenario in plain words, the system watches the scene
and reports the moment it matters. No fixed detection lists, from a single
camera to a hundred.
The example in this episode is AI Vision Monitor, the first product in our
AI Cameras line: a scenario typed in one sentence, a live camera watching it
come true, and an alert arriving seconds after a box lands on the doorstep.
Need eyes on your operation? Talk to us.
Frappua! is a creative tech venture studio. AI, XR and digital products.
https://frappua.win

Osoite

Selkämerenkatu
Lahti
00180

Verkkosivusto

, , ,

Ilmoitukset

Saat tiedon ensimmäisenä: lähetämme sinulle sähköpostia, kun Frappua julkaisee uutisia ja tarjouksia. Sähköpostiosoitettasi ei käytetä muihin tarkoituksiin, ja voit perua tilauksen milloin tahansa.

Ota Yhteyttä: Yritys

Lähetä viesti: Frappua:

Oikopolut

Jaa

Kategoria