plugnpl.ai is an Enterprise AI Execution Platform by just experts GmbH, Düsseldorf: a finished platform for AI assistants, automations and multi-agent workflows that runs inside the company's own Microsoft Azure tenant. Companies get the speed of a product and the control of building it themselves: full code access, free choice of models and each person's rights via Entra.

Where does plugnpl.ai run and where is our data stored?

plugnpl.ai runs in your own Microsoft Azure tenant, not in a vendor cloud. Sign-in runs through your Entra ID, secrets sit in your Key Vault, documents and histories in your storage, and we do not see your data. AWS and Google Cloud deployments are also supported.

How is plugnpl.ai different from Microsoft Copilot?

Microsoft Copilot is an assistant for working in Microsoft 365. plugnpl.ai is a platform for your own process-specific AI assistants and workflows, with full code access and free choice of models, run in your own tenant. The two combine well: Copilot for office work, plugnpl.ai for your own processes.

Why not just build it ourselves?

A custom build gives full control, but takes long and ties up a development team of its own. plugnpl.ai gives the same control, code in your own tenant and freely extensible, on a finished platform that is technically ready in days. Your team then only builds what makes your company different.

Which AI models are supported?

plugnpl.ai is model-agnostic: OpenAI, Azure OpenAI, Anthropic (Claude), Google (Gemini) and open-source models are supported, as are your own fine-tuned models. You choose the right model per use case and switch or combine providers without rebuilding the applications.

How fast can we go live?

Technically, plugnpl.ai runs in your tenant after a few days, because the platform is finished and is rolled out via Infrastructure-as-Code. Measurable benefit comes from the first use case once it is used day to day. That is why we start with a proof of concept, with the baseline, target and measurement period agreed beforehand.

Is plugnpl.ai GDPR and EU AI Act compliant?

plugnpl.ai is built for European regulation: hosting in your EU region, traceable data flows, a complete audit trail of every AI interaction and human approval steps in automated workflows. The legal assessment of each use case stays with you; we provide the technical documentation for it.

What happens if we stop working with you?

The platform keeps running in your tenant and you have access to the codebase. Your team, or any other Azure team, can operate and evolve it independently, and the models can be switched per agent.

The Enterprise AI Execution Platform · in your own cloud

The enterprise AI platform that works for everyone

For CIOs, CISOs and IT leaders: a governed platform to deploy AI assistants, automations and multi-agent workflows across the entire workforce, inside your own Microsoft Azure tenant. Full data sovereignty, complete code transparency, the rights of each person via Entra.

Book a free consultation Try the demo

Consultation: pick a slot with our team in the calendar. Demo: request trial access and try the platform yourself.

Runs in your own cloudMulti-LLM, open standardsGDPR & EU AI Act readyMade in Germany

Trusted in manufacturing, retail, logistics & engineering across the DACH region

4,6

How it works

How plugnpl.ai works

Agents work directly on the systems you already have. The organisation works with them through an interaction layer, people stay part of the process. just experts helps set up the processes.

All in your tenant, models in the EU. Consulting comes from just experts or, as white label, from a partner consultancy.

Product tour

The platform in eight steps

Let the tour play on its own or click through yourself. Every feature comes with what it does for the business.

1 / 8

AI Chat Interface

An intelligent assistant that writes content, analyzes data, develops ideas, and solves complex tasks, powered by the world's leading AI models. OpenAI, Claude, Gemini, DeepSeek: choose the best model for every task without switching providers. One login, one interface. No AI expertise required.

Business value: One access point for every employee instead of scattered AI subscriptions, with the right model for each task.

See it in action

From chat to Work Support

Your controls. Work across your systems.

Office suites reach far into everyday work, in the vendor's cloud and under the vendor's rules. Building it yourself gives control, and a long project. plugnpl.ai combines control in your tenant with work across your systems. Work Support takes it further: you hand over a task, plugnpl.ai prepares, you confirm.

AutomationHybrid automation platformse.g. IBM watsonx Orchestrate
Cloud agentsAgent platforms of the cloudse.g. Copilot Studio, Bedrock
DIY on AzureBuild it yourself on Azuree.g. Azure AI Foundry
ConsultanciesConsulting platforms in your tenante.g. MindsMachines
Open sourceSelf-hosted open-source chate.g. LibreChat
Sovereign AISovereign model platformse.g. Aleph Alpha
EU modelsEuropean model vendorse.g. Mistral
SearchEnterprise search with agentse.g. Glean
EU chatEU chat platformse.g. Langdock
Suite + work modeSuite assistants with work modese.g. Microsoft 365 Copilot
Cloud assistantsCloud work assistantse.g. Amazon Quick
Suite assistantsSuite assistantse.g. Gemini Enterprise
Model workspacesModel vendors' workspacese.g. ChatGPT, Claude Enterprise
Business suitesAgents in business suitese.g. ServiceNow, SAP
Chat toolsSimple chat toolse.g. nele.ai

plugnpl.aitoday

Work Support · end of 2026with Work Supportend of 2026

Categories with examples

Hybrid automation platforms e.g. IBM watsonx Orchestrate
Agent platforms of the clouds e.g. Copilot Studio, Bedrock
Build it yourself on Azure e.g. Azure AI Foundry
Consulting platforms in your tenant e.g. MindsMachines
Self-hosted open-source chat e.g. LibreChat
Sovereign model platforms e.g. Aleph Alpha
European model vendors e.g. Mistral
Enterprise search with agents e.g. Glean
EU chat platforms e.g. Langdock
Suite assistants with work modes e.g. Microsoft 365 Copilot
Cloud work assistants e.g. Amazon Quick
Suite assistants e.g. Gemini Enterprise
Model vendors' workspaces e.g. ChatGPT, Claude Enterprise
Agents in business suites e.g. ServiceNow, SAP
Simple chat tools e.g. nele.ai

Our reading, November 2026. The vendors stand as examples of their category.

Work Support arrives step by step.

Each step is something you can check: in the code, in the configuration and in the record of every run.

Today
Assistants and agents in your tenant
- each person's rights via Entra
- every change proposed, then confirmed
- scheduled tasks report back in a chat
November
Brakes in code
- scheduled runs read and draft, never carry out
- schedules from outside content wait for you
- a check after the answer
December
Every action on record
- a risk level on every confirmation
- "Show steps" for every run
- a cost cap per task
End of 2026
Work Support
- you hand over a task
- plugnpl.ai plans, works across your systems and prepares
- you confirm; every step is on record
Early 2027
Longer tasks
- tasks that resume after an interruption
- your rules block risky actions
- an error budget per agent; local folders

Plan as of November 2026.

Three ways compared

Consulting and platform from one source

What you get on each way. plugnpl.ai comes with consulting from just experts; as white label, other consultancies can offer it too.

Classic project

To first use: weeks to months
What stays: one project
Systems: what the project builds
Critical data: depends on the project
Cost: day rates around €2,000

Office suite assistant

To first use: days, in the vendor's cloud
What stays: the subscription
Systems: mainly its own suite
Critical data: under the vendor's rules
Cost: per seat

plugnpl.ai with consulting

To first use: days, in your tenant
What stays: a platform that keeps improving
Systems: Microsoft 365, ERP, business systems, own apps
Critical data: each person's rights, EU models, confirmation before every change
Cost: platform fee plus capped use; consulting as needed

Aspect	Classic project	Office suite assistant	plugnpl.ai with consulting
To first use	weeks to months	days, in the vendor's cloud	days, in your tenant
What stays	one project	the subscription	a platform that keeps improving
Systems	what the project builds	mainly its own suite	Microsoft 365, ERP, business systems, own apps
Critical data	depends on the project	under the vendor's rules	each person's rights, EU models, confirmation before every change
Cost	day rates around €2,000	per seat	platform fee plus capped use; consulting as needed

Checked every night

Every night, test accounts check whether the answers are right.

At a group of companies with 38 sites, twelve test accounts, one per department, ask the company chat the same questions every night, each with that department's rights. A judge compares every answer with the expected one. What fails becomes the next fix, and the next night shows whether it worked.

Every night

no manual spot checks: the test runs on a schedule, like a build

1,431

questions per night, from twelve test accounts with their own rights

< 10 %

wrong or invented in each of the last 3 night runs

1 · Asktwelve test accounts, each with its department's rights
2 · Judgeevery answer against the expected one: right, wrong, invented, no answer
3 · Fixin the knowledge graph or in the chat's instructions
4 · Next nightthe same questions again; a fix counts only if the numbers move