Aquarela · Success Stories · Cultura Inglesa
Success StoryCultura Inglesa centralizes access to generative AI with the AI Hub
From fragmented access to a single corporate platform for secure, centralized and governed use of generative AI models.
70
Active users on the platform
11
AI models made available
190M
Tokens processed
2K
Messages / interactions logged
Cultura Inglesa
Cultura Inglesa is one of the most recognized language-teaching institutions in Brazil, with a strong tradition in English-language education and a focus on educational innovation, student experience and the modernization of its internal processes. As generative AI advanced, different teams began to identify opportunities to use language models to support corporate, operational, pedagogical and administrative activities.
Fragmented AI access with no governance
Before the project, language models were used in a decentralized way, making it harder to govern, manage users, control consumption and standardize the experience across teams.
Isolated, fragmented AI accounts
Language models were used in a decentralized way, with different accounts, logins and isolated experiments across teams — creating a risk of duplicate contracts and costs.
Lack of governance and visibility
There was no single corporate entry point for generative AI, nor a standardized experience. It was hard to see who used which tools and to track consumption, adoption and model usage.
Complex onboarding and authentication
There was no authentication integrated with the existing corporate environment, nor a flexible foundation to add new models and providers in the future — which made secure, standardized adoption difficult.
AI Hub — a corporate generative AI platform
Aquarela designed and implemented the AI Hub, the central point of access for Cultura Inglesa users to the main language models — removing the need for individual or isolated accounts across different providers.
01
A single entry point for AI
The AI Hub became the central environment for accessing the main language models, removing the need for individual or isolated accounts across different providers. Users access everything through one corporate interface.
02
Login with the corporate email
Access was integrated with Cultura Inglesa's corporate email, connected to the company's Google domain. This simplified onboarding, lowered adoption barriers and strengthened usage traceability.
03
Multiple models through a proxy layer
Using LiteLLM as a routing and standardization layer, the platform unifies AWS-native models (Amazon Bedrock) and external providers — combining different model families in a single experience.
Models made available
In total, 11 models in a single platform — users pick the one best suited to each need.
Anthropic
OpenAI
Google Gemini
A single, centralized and governed platform
In operation since July 2026, the AI Hub reduced access fragmentation, simplified onboarding and created a secure, flexible foundation to expand AI usage across the organization.
Centralized, governed access
Multiple AI models (Anthropic, OpenAI and Google Gemini) brought together in a single platform, reducing the fragmentation of isolated accounts and bringing more governance over generative AI usage.
Simple and secure onboarding
With login through the corporate email integrated with the Google domain, adoption became simpler and authentication more secure — with usage traceability per user and per model.
A scalable foundation for the future
The flexible architecture allows new models to be added without restructuring the environment, creating a solid base for new generative AI initiatives across the organization.
Technologies used
Want results like these at your company?
Talk to our specialists and discover how to centralize and govern generative AI usage across your organization with Cloud, DevOps and agentic AI on AWS.
Talk to a specialistOr visit aquarelacs.com.br to learn more about Aquarela.