fish in a bottle: Transforming Children’s Language Learning with AI-Powered Personalisation

fish in a bottle is an award-winning UK educational games company that joined cohort 8 to develop its latest AI-powered language learning project, Navigo.

By combining pedagogy, game design and emerging technologies, award-winning UK educational games studio, fish in a bottle, is on a mission to make learning more engaging, inclusive and effective. Specialising in curriculum-aligned digital learning experiences for children, the company has a strong track record of delivering products at scale for both national and international partners.

The team joined Cohort 8 of the CDI Impact Accelerator to progress their latest project, Navigo, an AI-powered English language learning experience for children aged 6-12 years aimed at solving the challenge of scalable personalisation. Through the programme’s support, fish in a bottle built a fully functional MVP, improved the scalability of its technical architecture, and refined its commercial strategy.

As a result, Navigo evolved from an early concept to a market-ready prototype, and the company successfully demonstrated that personalised, culturally adaptable language learning can be delivered at scale.

The Challenge

High-quality, scalable personalisation in English language learning remains a major challenge, especially for international audiences. Existing solutions often rely on static content, limited adaptivity, or teacher-intensive authoring workflows, making it difficult to tailor learning experiences to each child’s linguistic level, cultural context, and motivational needs.

Early research conducted by fish in a bottle indicated that narrative engagement, meaningful choice, and scaffolded challenges significantly improve motivation and retention for language learners. However, turning these insights into practical digital products has historically been constrained by the cost and complexity of content creation, and the difficulty of deploying AI safely in child-facing contexts. Navigo was conceived to directly address this gap, bridging the space between cutting-edge research and real-world classroom-ready technology.

The Solution

Navigo combines AI‑assisted authoring with adaptive narrative gameplay to enable scalable personalisation without compromising pedagogical quality or safety. Its technical solution is built around two core components: an AI‑Assisted Authoring Tool for educators and designers, and a runtime narrative engine that delivers personalised, interactive learning experiences for children.

AI‑Assisted Authoring Tool

The authoring tool provides educators and designers with an AI‑enhanced workspace that blends controlled LLM‑generated content with established pedagogical principles. This allows for the rapid creation of culturally adaptable narratives and finely calibrated language challenges that can be deployed directly into production workflows.

Adaptive Narrative Engine

Complementing the authoring environment, the runtime narrative engine personalises the learning experience in real time. By adapting story pathways, challenge difficulty, and feedback, it ensures each learner receives meaningful support grounded in rule‑based and probabilistic instructional logic.

“Technical decisions prioritised transparency, human‑in‑the‑loop design and separation of concerns,” says fish in a bottle Founder Drew Wilkins, “ensuring that AI supports—rather than replaces—expert judgement.” This approach distinguishes Navigo by embedding research‑driven pedagogy directly into both the authoring and runtime systems, translating academic insight into tools that support real production pipelines and classroom‑ready products.

Accelerating Development with the UCL CDI

Participation in the CDI Impact Accelerator significantly accelerated both the technical and commercial development of Navigo. With access to focused technical mentoring, structured validation, and expert input, the team moved rapidly from concept-level exploration to a demonstrable MVP within weeks.

Technical Achievements

• Delivery of a functional MVP of the AI-Assisted Authoring Tool, demonstrating safe and controllable use of large language models for narrative and language content creation.
• Establishment of a scalable technical architecture capable of supporting future market expansion.
• Refinement of adaptive logic linking learner performance to narrative progression.
• Implementation of early data structures for learner analytics.

Together, these milestones reduced technical risk and validated the feasibility of Navigo beyond a research prototype.

Business Achievements

In parallel, the programme enabled significant commercial progress. The team sharpened Navigo’s product scope, value proposition, and commercial objectives, resulting in greater clarity around its planned route to market in October 2026. With ongoing input from experts at the CDI and the wider UCL ecosystem, the team accelerated market research and evidence-based decision making, particularly in relation to international markets. As a result, Navigo’s product-market fit has been refined, increasing confidence from partners and stakeholders, as well as strengthening readiness for future investment.

Looking Ahead: Preparing for International Launch and Long-Term Scalability

Following the CDI programme, fish in a bottle is focusing its technical development on optimising LLM integration to achieve lower latency, support bulk processing, improve system prompts, strengthen security, and enhance scale management. As Navigo progresses toward market readiness, the team will also build robust production environments, shifting from rapid MVP iteration to long-term, maintainable infrastructure.

Additionally, the company plans to broaden its use of LLMs to generate additional educational content across its wider product portfolio, while expanding the adoption of AI‑assisted development tools within the organisation.

From a business perspective, the immediate priority is to fully integrate AI-assisted authoring into Navigo, positioning it as a core feature of the game. Market entry will initially target Mexico as a strategic gateway to broader distribution across South America, working with regional partners to ensure cultural and curricular alignment. Alongside this, fish in a bottle will continue developing the authoring capability as a reusable asset, supporting future learning products and enabling scalable, partnership‑driven growth.

To learn more about fish in a bottle, visit: https://www.fishinabottle.com/