LEAP 2026 · Riyadh · 31 August to 3 September · Stand H1A.P93

Every call, in every language. Just works.

Real-time translation inside phone calls, voice agents and live rooms, without changing the systems you already run.

One problem, three places it shows up

Customer service

The customer speaks Arabic. The agent speaks English. Both keep speaking naturally and the call is translated live, on the number the customer already called.

Voice agents

The AI handles the call in the customer's language. When a human takes over, the language and the context go with them.

Live events

One speaker, an audience speaking many languages. Attendees scan a QR code, pick a language and listen on their own phone.

These are Translate, Talk and Events. One platform underneath.

Why SentiVue

Invisible

Works inside the phone call and the voice infrastructure you already run. Nothing for the caller to download, no change to your call flow. If translation is unavailable, the call still connects.

Translation survives the handoff

Most voice AI stops the moment a human picks up. Ours does not. The caller never has to switch language halfway through.

Built for dialects

We train our own speech models on native speech data for the language variants that general-purpose models handle badly.

Runs where you need it

Cloud, in-Kingdom, or fully on-premise inside your own data centre.

Try it yourself

The Events surface takes about two minutes to run with two phones. Tell us and we will set up a room for you.

1We send you a room link. Open it on your phone and pick the language you will speak.
2Send the join link to someone else, or let them scan the QR code. They pick their language.
3Speak normally. They hear you in their language, and you hear them in yours.

Before you try it

Use earbuds or headphones. Two phones playing out loud in the same room will hear each other and loop.
Speak in full sentences. Single words give the system less to work with.
Expect a short delay. It feels like speaking through an interpreter rather than a phone call. That is the nature of live translation, not a fault.
A quiet room helps. Background noise is the hardest condition for any speech system, and an exhibition hall is the hardest room there is. Come to stand H1A.P93 and we will hand you headphones and a quiet corner.

What is proven so far

1.6 s

translation pipeline

70+

languages

Own voice models

built for real dialects

Selected for Vodafone's Open Innovation Program 2.0 and the NOS 5G TestBed.

Delivered live translation at the 2025 Cybersecurity Conference in Mannheim, for organiser SAMA Partners.

European Portuguese and Egyptian Arabic are trained and running in production. Saudi dialects are in progress, with Najdi and Hijazi as the current research focus. We do not show them as available, because they are not yet.

Where it runs

Cloud, in-Kingdom, or fully on-premise inside your own data centre. We have run the full pipeline on a customer's own hardware, so voice data does not have to leave the environment you choose.

If you answer to a regulator about where calls are processed, that is a design decision we have already made, not a roadmap item.

What we're looking for at LEAP

We are looking for Saudi organisations with a real language problem in customer conversations, who want to test SentiVue in a defined environment.

Contact centres serving customers across languages
Enterprises deploying voice agents
Telecom operators exploring multilingual calling
Organisations with data-sovereignty or private-deployment requirements
Events that need live multilingual access
Systems integrators and technology partners

Investing?

We are happy to talk. Tell us what you would like to discuss.

Register your interest

Tell us what you need and we will come back with something specific rather than a brochure. About forty seconds.

Find us at LEAP

Stand H1A.P93, 31 August to 3 September.

Ask us for a live call and we will put you on the phone with someone speaking a language you do not.

Rocket Fuel Stage, Hall 1A

Round 3, Tuesday 1 September, 1:30 to 4:00 PM

We are on at approximately 2:34 PM