LIVRESQ Companion — AI teacher assistant launched as a Google for Startups result

LIVRESQ Companion sits at the intersection of three records that buyers rarely see colocated: a Google AI-for-public-sector programme participation, an NVIDIA Connect ISV partnership, and a BVB-disclosed first contract with named teacher and school counts. The three records together place the product on procurement-grade evidence — not pre-product positioning, not a self-reported launch, but a sequence of independently-verifiable institutional milestones.

For ministries and education-system buyers evaluating AI-assisted teaching tools, the relevant question is not "does the model work" but "is the product institutionally legitimate". LIVRESQ Companion answers that question in three places: a Google global accelerator selected it as the only Romanian entry in its EMEA cohort; NVIDIA accepted Ascendia into its ISV programme; the Bucharest Stock Exchange holds the disclosure of the first contract. None of these records substitute for product evaluation; together they substitute for institutional risk.

The Companion is built for teacher productivity inside the LIVRESQ authoring environment — not as a general-model chatbot. That positioning matters because the procurement frame for AI-in-public-sector tools is typically not "general AI" but "AI for a specific public-sector workflow". The Companion answers that frame for the lesson-authoring workflow.

The first LIVRESQ Companion contract reaches 10,000+ teachers across 800 schools in Romania. The contract was disclosed to the Bucharest Stock Exchange on 8 April 2026 under current-report ASC_20260408. The contract is the BVB-disclosed launch of the product line; subsequent contracts will be disclosed under the same current-report framework as they cross the BVB-disclosure threshold.

The product surface is integrated into the LIVRESQ authoring environment that 158,900 creators across 7,260+ institutions already use. Teachers do not need to switch to a separate AI tool — the Companion sits inside the existing authoring workflow as a productivity layer. The integration pattern is the same one Ascendia uses across the platform: AI as workflow capacity rather than as a separate destination product.

The first contract's 10,000+/800 figures place LIVRESQ Companion among the largest single-contract teacher-AI deployments publicly disclosed in the EU at the time of the BVB filing. The procurement-grade nature of the disclosure (current report, not press release) is what distinguishes the launch from the ambient AI-in-education noise.

Google for Startups Growth Academy: AI for GovTech is a global accelerator for AI-based public-sector solutions. The first EMEA cohort selected 25 solutions globally; LIVRESQ was the only one from Romania. The three-month programme provided mentorship from Google executives focused on AI-for-public-sector applications and accelerated product development of the Companion.

The selection is independently verifiable on the Google for Startups public record. For procurement evaluators, this is the kind of authority signal that distinguishes a product launched on the back of a global programme review from a product launched on internal positioning. The Companion's entire product origin story is on a public Google programme record.

LIVRESQ Companion is operated under listed-company governance from EU-resident infrastructure. The R&D and AI-governance frame is owned by Gabriel Lazar, Ascendia's Research & Development Director and the named member in the European Commission Digital Education Content (DEC) expert group since 1 July 2024. EC DEC carries the EU-level guidelines work on quality requirements for digital learning materials, including AI-mediated content. Having that membership inside the Ascendia leadership surface places the Companion under a published institutional AI-governance reference rather than vendor-level attestations alone.

The procurement-relevant claim is institutional, not model-specific: the product is delivered by a listed company, with a named EC DEC expert-group lead carrying the AI-governance frame, under BVB-disclosed contracting, on EU-resident infrastructure. For ministry buyers comparing AI-in-education products, that combination is the evaluation criteria stack — model performance is necessary but it is not the dimension that decides procurement.