Ridenn Zavatta

AI agents · Automation · Business tools

AI agents.
Tools that serve a purpose.

I design AI agents and custom tools to connect your applications, make use of your documents and simplify your teams’ work.

Discuss your projectSee the demo

From audit to go-live, then onboarding for your teams.

  • Cloud & APIs
  • Business applications
  • Sovereign AI
Explore a project

01A project in production

From need to product.

Plaidoria reference check: two citations verified against official sources, and one reference that could not be found, struck through and marked “génération bloquée” (generation blocked).Plaidoria · Product in production

Plaidoria

A legal drafting tool I designed, built and run myself. The references it produces are checked against official sources, and the lawyer stays in control of the document.

Use
Sourced legal drafting
Role
Design, development, operation
Status
Commercial product in production
Take the guided tourExplore the Plaidoria project

02What we can build

The right system, for the right use.

Connect your tools, process your documents or build a business application. AI comes in where it adds value; for established rules, deterministic processing is enough.

  1. AI agents & API integrations

    Agents · APIs · Automation

    Connect an assistant to your tools to search, prepare or carry out actions, with explicit permissions and controls.

  2. Document automation

    Transcription · Extraction · Summarization

    Turn recordings and documents into structured information, with the necessary checks.

  3. Applications & business tools

    Internal applications · Interfaces · Workflows

    Build an application that fits your processes, from the interface to the connections with your existing tools.

  4. Sovereign AI & knowledge

    Local LLMs · RAG · On-premise deployment

    Put your documents to work, integrate a local model and keep control of data flows when your context requires it.

Explore the expertise

03Engagement formats

Three ways to move forward.

Clarify a decision, build a tool or support your team. Where you start depends on what you need today.

  1. Audit & scoping

    Know what to build, with which data and for what use.

    See this format: Audit & scoping

  2. Design & deployment

    Move from a defined need to a usable, verified system.

    See this format: Design & deployment

  3. Training & support

    Help your teams choose, use and evolve their tools.

    See this format: Training & support

04Selected work

Different problems. The same standard.

Internal applications, automation and AI: each project details its context, its constraints and the solution built.

Editorial cover for the project: the signature word “Flux” (French for flow) as a serif watermark and a line-art pattern of ticket queues, cobalt ink on a dark background.

Sovereignty · Business app

Sovereign queue management

A business application designed to replace a proprietary solution and take back control of the visitor reception tool, how it is deployed and how it evolves.

What the project demonstrates

Sovereignty, streamlining, reduced dependency.

Read the case study
Editorial cover for the project: the signature word “Voix” (French for voice) as a serif watermark and a line-art pattern of a sound wave turning into lines of text, cyan ink on a dark background.

Local AI · Automation

Transcription & AI reporting

A local transcription and summarization workflow designed to reduce the documentation workload and speed up operational processing.

What the project demonstrates

Automation, productivity, time saved.

Read the case study
Editorial cover for the project: the signature word “Salon” as a serif watermark and a line-art concierge seal, gold ink on a dark burgundy background.

Agentic AI · Full-stack

Le Salon

A full-stack platform that covers the whole meeting cycle: booking, an AI Concierge that takes action (tool calling), a real-time SIDSIC Cockpit for technicians, and minutes generated from a simple audio file. A deliberate palace-style identity, with an optional 100% offline mode.

What the project demonstrates

The full meeting cycle, AI that acts instead of just answering, product-grade finish on an internal tool.

Read the case study
All case studies

Starting points

Before we start.

The choice of solution comes after understanding your need.

My need is still vague. Where do I start?

Describe a task that takes too long, a missing tool or a decision to make. The first conversation clarifies the problem; an audit can then help define a scope before any development.

Do I have to use AI?

No. A business rule, a connection between two applications or conventional software can be enough. AI is chosen when it brings real value to the use case, with results and limits that can be verified.

Cloud or on-premise deployment?

Both are possible, depending on the project. The choice depends on data sensitivity, your tools, the expected performance and your operating resources. These constraints are clarified during scoping.

How do we prepare for what comes after delivery?

Documentation, onboarding, operations and future changes are planned from scoping onward. What will be handed over and the support you expect are set out in the scope of the engagement.

Let’s build what comes next

Let’s talk about your next tool.

You deal directly with Ridenn Zavatta.

A few lines about your context and what you want to improve are enough to get started.

  • Your context
  • The outcome you want
  • Your known constraints
Discuss your project

The first conversation

We clarify the need and the first useful step: scope an idea, build a tool or support your team.

Or directly by email
zavattaridenn@gmail.com