Technologiczna organizacja non-profitOpen-sourceStworzone dla realnego wpływu

AI to empower people in need
and those who help them

The first open-source AI agent platform built for public services.

Gates Foundation
Google.org
Fast Forward
Raise Sherpas
L'Assurance Maladie
France Travail
Hôtel-Dieu (AP-HP)
BOpEx (AP-HP)
AP-HP
La France s'engage
California Department of Justice
US Dept of Veterans Affairs
Y Combinator
Gates Foundation
Google.org
Fast Forward
Raise Sherpas
L'Assurance Maladie
France Travail
Hôtel-Dieu (AP-HP)
BOpEx (AP-HP)
AP-HP
La France s'engage
California Department of Justice
US Dept of Veterans Affairs
Y Combinator
Nasze rozwiązanie

Bezpieczna, skuteczna i suwerenna SI dla usług publicznych

1Bezpieczna

Stosuje się do Twoich procedur zamiast do nawyków modelu, podaje źródło każdej informacji, a Twoi eksperci zatwierdzają ją przed wdrożeniem.

2Skuteczna

Mierzy się ją w działaniu, po wynikach w terenie: czas odzyskany przez zespoły, szybsze odpowiedzi dla obywateli, zgodność monitorowana na bieżąco.

3Suwerenna

Działa na Twojej infrastrukturze, z Twoimi danymi, jako open source. W każdej chwili możesz przejąć nad nią kontrolę.

Problem

Generyczna SI nie powstała z myślą o usługach publicznych i społecznych

Miliony osób każdego dnia są zależne od pracowników w terenie. A komercyjna SI nie została zaprojektowana z myślą o nich.

Nie zna ich procedur.
Nie chroni danych wrażliwych.
Halucynuje wtedy, gdy ma to największe znaczenie.

Kim jesteśmy

Wiodąca organizacja non-profit wykorzystująca sztuczną inteligencję, by od nowa kształtować usługi publiczne na masową skalę

Założeni w 2014 roku w Dolinie Krzemowej, dostarczamy otwarte rozwiązania oparte na SI, którym zaufały rządy, organizacje międzynarodowe i lokalne społeczności w ośmiu krajach, służąc milionom obywateli.

Dowiedz się więcej
12+
lat budowania SI dla dobra wspólnego
8
krajów
20M+
obywateli korzystających z naszych platform
Our product

Introducing the Bayes Platform

The first open-source AI agent platform built for public services. Here is what it does, in the real product.

Open the platform tour
Bayes Platform
Najedź na kartę, by dowiedzieć się więcej.
01
Grounded in your own corpus
Grounded in your own corpus

The agent answers from your handbooks, protocols and official references: resource libraries, tagged documents, PDF import, a site indexed by crawl. Never from the open web.

02
Every answer cites its source
Every answer cites its source

The sources panel shows the document title and the extract used, so whoever reads the answer can check it. An answer nobody can verify has no place in a public service.

03
Your protocol, held by the platform
Your protocol, held by the platform

The agent's settings are your organisation's. Every reply has to report its own bookkeeping: session title, categories drawn from your list, sources cited. The platform enforces that when the model skips it.

04
Draft, review, publish
Draft, review, publish

Every change writes a draft. Publishing makes it the version that runs, named and dated in the history, and every reply carries its version badge. Nothing reaches the public that a human has not published.

06
Tested before it goes live
Tested before it goes live

Evaluation campaigns replay conversation and extraction datasets against a chosen version. Reviewers rate the answers blind, an LLM judge gives the scale, and two runs compare side by side.

07
Watched in production
Watched in production

Every exchange is filed under the life situations you defined. The topics your answers cover badly surface on their own, instead of waiting for a complaint.

08
It plugs into what you already run
It plugs into what you already run

Each agent exposes itself as an MCP server, so the assistant or portal your institution already runs calls it as a tool. A widget on your site or a public page when that is simpler. Nobody changes screen.

09
Runs where you decide, open-source
Runs where you decide, open-source

Managed cloud, or your own infrastructure from a single Docker image. Document isolation is written into the retrieval query itself rather than a checkbox, the retention period is yours, and the code is MIT-licensed.

Health · MCP grounding

AI grounded in your own protocols.

A general-purpose assistant answers plausibly but sometimes wrongly; grounded in your official, cited guidelines, it answers correctly.

Demo

The same clinical question, put at once to a generic assistant and to the MCP-verified platform.

Employment · France Travail

A proactive AI work coach.

An agent that draws on the platform tools to answer, and stays within its remit: out of scope, it routes you to a human counsellor.

Demo

A coach that calls tools, surfaces openings and benefits, and hands off to a counsellor when a request is out of scope.

Public services · Togo

A whole public service, by chat.

A full administrative procedure handled end-to-end on WhatsApp, without leaving the conversation.

Demo

A diploma request: identity (e-ID), auto-fill from the official records, and payment.

Discover the Bayes Platform
Care plan for a 4-month-old with mild acute bronchiolitis?
● General-purpose AINo MCP
    ● Bayes · verified vs HASWith MCP