AI is already in your company. Is it under control?
Companies are adopting AI fast, but the value and the control often lag behind. Employees use personal accounts when the company’s tools don’t meet their needs, and teams buy separate licences and API keys. Nobody has a clear view of total spend, of which data leaves the organisation, or of which agent is calling which system.
At Code10 TechTalks 2026, Fernando Cucci, Solution Architect at Code10, presented the product built to answer those questions: the C10 Control Tower.
One gateway for every AI call
The idea is simple. People, applications and agents sit on one side, and AI models on the other. The Control Tower sits in the middle, built on Amazon Bedrock with access to more than 250 models, and every request passes through four steps:
- Identity check: confirm who is asking and whether they can use that model
- Data protection: block sensitive prompts, or anonymise personal data before it reaches the model
- Smart routing: send simple questions to cheaper models and complex ones to more capable models
- Full logging: record cost, tokens, latency, user and model for every call
What you will see in the demo
Fernando walks through the user interface and the administration console. He shows shared knowledge bases, reusable projects that combine prompts, MCP tools and documents, and guardrails that stop a request before it costs anything. He also shows an auditor workflow designed for GDPR, where every review requires a stated reason and is itself recorded.
The same rules apply whether requests come from the web interface, Office plugins, Claude Code, browser extensions or your own applications. Prompt compression helps cut costs further.
Talk to our experts
Want to bring visibility and governance to AI in your organisation? Contact us to see the C10 Control Tower in action.
