just AI
plugnpl.ai the concept behind the support hub
The support hub is based on the system developed by just experts developed by plugnpl.ai-concept developed by just experts. The basic idea is is to maximize flexibility in the integration of any data sources, AI models and functions. functions and to be able to use these directlyConnect and get started.
The plugnpl.ai concept and the requirementsrequirements and goals of modern software development are the basis ofs supporthubswhich was developed by just experts developeddeveloped AI platform.
Requirements and goals
The 3 principles of plugnpl.ai are Flexibility, agility and speed.
Therefore the platform meets the following requirements:
- The core application is regularly maintained and centrally updated. Essential updates are developed centrally and can be rolled out in different environments and to different organizations.
- Enable custom development
- Each company can develop its own AI-based components that can be easily integrated into the platform. This allows plugnpl.ai to be customized according to individual needs. These individual developments can be shared and expanded with other developers as part of the Supporthub community
AI enables it people people to work more efficiently.n. The basis for this are all applications are immediately available and have an appropriate speed, without delay.
Stateless Components are elements that do not save or change any internal data. They function solely on the basis of the information passed to them and always deliver the same result for the same input. As a result, they are generally easier to understand and test and ensure better performance.
Further developments and updates from plugnpl.ai can be "one-click" can be rolled out. No time-consuming and manual deployments.
As soon as there are new services and functions in the infrastructure, the dthese immediately integrated into your infrastructure components. This avoids avoids technical complications and enables rapid further developments.
Flexibility makes uour platform. Individual features or specific functions can be integrated with little effort, so you can, for example, connect your ticket system to the platform or have pre-trained LLMs detect possible quality problems in production. As long as you can imagine it, we can also implement itthere are no limits!
Safety is our top priority. Dhe software was from the very beginningin all elements secure and integrated designed. Secure code development, continuous and integrated authentication - interactive and non-interactive
In the Supporthub interface we integrate Your logo and your brand colors according to your ideas, to seamlessly integrate the solution into your CI and your systems. anadd. We also offer the integration of a company profile and industriespecific prompt properties can also be seamlessly integrated.
Do you knowwhat the nextn bign AItrends will be? No? We either! But we are ready, these to integrate them into our platformwhatever it may be.
The operating model of the support hub is tailor-made for your organization and can be customized for every need and scaling can be customized: From your code on theiner infrastructure all the way to a lean SaaS model.
Developers should focus on developing and their time not with manual integration and deployment taskscomplex configuration or network problems waste. Automation, standardization and a high level of integrability reduce this task to an absolute minimum.
10 users or 10,000 users per day, the performance of the platform remains the same.
Automate everything that needs to be done more than once. Repetitive tasks or process steps are taken over by the platform.
Long wwait until Dan answer has been completely generated by the LLM? Not on the plugnpl.ai platform - All responses are generated across devices and sessions immediately and behave synchronously. Sending a request from your cell phone and laptop at the same time? No problem - your requests are processed simultaneously!
Multi Layer Architecture
A multi-layer architecture is a comprehensive system design that includes different layers to handle specific tasks and create an efficient, flexible and easy-to-maintain infrastructure.
This level refers to the people who interact with the system. It includes:
- Comprehensive Training Concept: A comprehensive training concept that ensures that all users are well prepared to use the systems effectively.
- Seamless SSO with Entra ID / Graph Integration: A seamless single sign-on (SSO) solution that simplifies access to the system and ensures that users only need to log in once to access all resources.
This layer contains the applications and services that end users use directly. These include
- Supporthub Web App: A web application that offers various support functions such as websearch, translate or reformulate.
- Customer Specific ERP Integration: Integration with specific ERP systems to optimize business processes.
- Power Automate Integration: Automation of workflows and processes through integration with Power Automate.
- Teams Client Integration: Integration with Microsoft Teams to improve collaboration and communication and to be able to use data and information from Teams.
- Edge integration: Enables applications to communicate directly with devices at the edge of the network, such as sensors or machines, and to process data in real time. This leads to faster responses and more efficient processes, as the data does not have to be sent through the entire network first.
This layer ensures that different systems can communicate seamlessly with each other. It includes:
- Supporthub API: An API that enables access to the functions of Supporthub.
- "Feature XY" API scope: Specific API scopes that enable the customization and extension of functions.
This layer is responsible for processing and analyzing data. This includes
- Optimized ChatGPT Connect: Optimized connections to ChatGPT to provide powerful AI features.
- Generic LLM Connect: Connections to generic Large Language Models (LLMs) for various use cases.
- Analytics Tool for SQL: Tools for analyzing SQL data.
- Websearch: Search functions on the web to connect the latest information.
- Sales Agent: AI-controlled sales agents that support the sales process.
This layer includes any Large Language Models (LLMs), which can be used for various tasks:
- ChatGPT on Azure: Provision of ChatGPT on the Azure platform.
- ChatGPT on OpenAI: Use of ChatGPT via the OpenAI platform.
- Self Hosted, Self Trained LLM: Own hosted and trained LLMs.
- Any 3rd party LLM API (such as groq (Llama 3) or Google Gemini): Use of 3rd party LLMs such as groq or Google.
This layer is responsible for storing and managing data:
Persistente datastorage (e.g. context or software use): Permanent storage of data relevant to the context and use.
This layer comprises the company's specific data structures:
This multi-layer architecture makes it possible to manage different tasks efficiently and respond flexibly to changing requirements. Each layer specializes in certain functions and helps to ensure that the overall system is robust, scalable and easy to maintain.
Operating Model
Our operating model offers flexible solutions that can be easily adapted to your individual needs and scaling requirements. It includes multiple migration paths and infrastructure options specifically tailored to different scenarios:
as SaaS
(Software as a Service)
- Own database, our Entra ID (B2C/B2B)
- Included services:
- Entra Identity
- Static apps
- Chat service
- Dedicated SQL database
- Shared environment, your Entra ID
- Included services:
- Entra Identity
- Static apps
- Chat service
- Dedicated SQL database
- Your own environment, your Entra ID
- Included services:
- Entra Identity
- Static apps
- Chat service
- Dedicated SQL database
or your infrastructure
- Own database, our Entra ID (B2C/B2B)
- Included services:
- Entra Identity
- Static apps
- Chat service
- Dedicated SQL database
- Own repositories, own customer branch (dev/stage/prod)
- Included services:
- Entra Identity
- Static apps
- Chat service
- Dedicated SQL database
Our model facilitates migration and offers seamless integration yourexisting infrastructure. No matter whether you prefer a shared environment or a dedicated environmentstwe have the right solution for you.
Design principles
The design principles of the architecture are aimed at maximizing flexibility and agility while minimizing risks.
Multi Layer Architecture
In a multi-layer software architecture, the system consists of different levels or layers (multi-layer architecture), each of which fulfills dedicated tasks. Each of these layers can be individually adapted, e.g. by adding customer-specific modules, without affecting the functionality of the entire system.
If the technological environment changes, the affected level is updated without affecting the other levels. This means that the system remains flexible and adaptable, even if the requirements or technologies change. Individual levels and functions can be integrated as required: different front end? Then connect directly to the integration layer; own LLM? Then this can be integrated into the LLM layer.
SaaS or Azure Subscription
We offer you the option of using our software either as SaaS (Software-as-a-service) or host the platform in your own Azure subscription. With the SaaS option, we take care of the entire infrastructure and maintenance so that you can without your own hosting access the software. When you host the platform in your own Azure subscription, you have control over the installation and management of the software in your own cloud environment. Regardless of which option you choose, all users can all users can receive regular updates receiveto ensure that the software is always up to date and offers the best possible performance.
All modular units
All components wwere as statelessindependent modulesto enable customization. Stateless means that the services do not store any data and can be used independently work independently. The modules are small, separate parts that can be easily adapted or separated separated. Everything is parameterized, which allows flexible settings to respond to different requirements to react to different requirements.
Entra ID integrated
Lightning fastit and safe Alog in via Entra ID (formerly Azure Active Directory) enables an integrated login via Single SignOn (SSO), incl. multifactor authentication if set up.
Graph API integrated
The Supporthub knows your Entra user - profile picture, display nameand group memberships are synchronized. A native display of the Teams including channels and much more is easily possible in the user context via the integration of the Microsoft Graph interface.
API contracts
In order to design robust and future-proof API contracts, clear and stable interfaces are defined that can be easily extended. This enables easy integration of new functions and extensions in both the backend and frontend without affecting existing implementations. In addition are detailed documentation and versioning systems systems are available to ensure smooth further development and and expansion of the use of the API in various places to guarantee.
Floating platform
If a part of the platform e.g. by one of ther hyperscaler easier and/or or better, this is not a problem: We utilize and integrate innovation instead of trying to outdo the leading innovators. Our focus is on providing the best solutions for our customers, regardless of who developed them. This approach allows us to always use the most advanced technologies and continuously improve our offering.
All Infrastructure as Code (IaC)
All infrastructure components are designed from the outset as "Infrastructure as Code" (IaC) to enable effortless deployment and scalinglization and possible. This means that every change to the infrastructure automated, consistent and traceable be carried out.
All Zero Trust
We implement Zero Trust design principles from the very first line of code. Zero Trust assumes that no one is trusted, whether inside or outside the network. Every request is checked and authenticated before access is granted.
Pay per usage
We develop all system components in such a way that they only incur costs when used. This reduces fixed costs and enables flexible, efficient expenditure.
Infrastructure
Our infrastructure ensures seamless integration and offers you a high-performanceflexible and scalable solution. The user interface starts with the user, who can use Static Apps (frontend),APIs and HTTP (HTTP, ) to access our services.
- The Entra ID manages authentication, while the API Management Service ensures that all API requests are handled efficiently and securely.
The chatservice enables real timekommunication and interacts via WebSockets with Web PubSub for the streamingand broadcasting. The data is stored securely in SQLserver and managed.
For the use of state-of-the-art AI technologies, we offer OpenAI on Azure, whichs via the Chat Completion API and enables streaming (SSE).
Our infrastructure also includes important services such as application insightsfor monitoring and analysis, Key Vault for secure key management and Graph API for complex data queries and interactions.
The entire infrastructure is managed via Terraform, enabling consistent and repeatable deployment.
This comprehensive architecture ensures that your applications are not only secure and reliable, but can also be adapted to new requirements at any time.
CI/CD
Continuous deployment and integration
We have built a comprehensive continuous integration and deployment process that enables lightning-fast deployments without downtime and supports the development of customized solutions. Our process is divided into two main areas:
- Development environment and
- Customer environment
Development environment
- Your developers write the code and check it into Azure Repos.
The code is stored securely in Azure Repos.
- The code runs through a series of tests:
- Linting
- Unit tests
- End-to-end tests
- PR review
- The checked code is processed further:
- Deployment in the staging environment
- Acceptance tests
- Deployment in the production environment
Customer environment
- You check your code into the customer repo.
The code also runs through tests (similar to the development environment).
The code is provided in different environments:
- Dev: Chat service, static apps, infrastructure via IaC/Terraform
- Stage: Chat service, static apps, infrastructure via IaC/Terraform
- Prod: Chat service, static apps, infrastructure via IaC/Terraform
- Prod Customer: Chat service, static apps, infrastructure via IaC/Terraform
Your customer environment can be operated either as SaaS (Software-as-a-Service) or via your own Azure subscription.
Development partnership
Let's start
together
Benefit from the achievements of the Plugnpl.ai community and become an active part of the future of AI!