# AI Integration Services Source: https://ordergroup.co/ai-software-development/ai-integration-services/ Last updated: 2026-10-22 > AI integration services for products that already work: a separate AI service, a swappable model, human checks and a free AI readiness assessment. # AI integration services: add AI to the product you already have A new AI feature must not break a product that works, and it must not show users content nobody has checked. We add AI to web and mobile products as a separate AI service that your application calls, with a person or a test checking the output. [Book a free consultation](https://ordergroup.co/contact-us/) Tell us what you're building. A few details help us bring the right engineers to the first call. - ![AASA24](https://ordergroup.co/media/images/aasa24-mono.max-208x64.format-webp.webp) - [![AP-Tech](https://ordergroup.co/media/images/ap-tech-mono_ihuJP5T_6b2rSSb.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/monster-airport-lighting-system/) - ![Avafin](https://ordergroup.co/media/images/avafin-mono.max-208x64.format-webp.webp) - [![BizBot](https://ordergroup.co/media/images/bizbot-mono_StdJIUN.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/bizbot-fintech-ux-design/) - [![Enerace](https://ordergroup.co/media/images/enerace-mono_FKuME1X_Fp3WF4B.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/enerace-energy-trading-software/) - ![Filmgrail](https://ordergroup.co/media/images/filmgrail-mono_gPdfbHO.max-208x64.format-webp.webp) - [![Kyoto Group](https://ordergroup.co/media/images/kyoto-group-mono_hlqECHG.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/kyoto-digital-twin-software/) ## AI integration services for products that already work AI integration services add an AI feature to a product that is already in use: a chat, a product-search agent, live speech-to-text or summaries of market news. We have added such features to Enerace, Centrum Komunikacji and a product platform from the renovation market. A fourth, Generator ETR for PSONI, an app that turns text into easy-to-read language, is in development. What we added to each, and how it connects ## Four integrations in live and in-progress products Enerace and Centrum Komunikacji got AI features inside platforms that were already running. The renovation platform shows the separate service, Generator ETR the swappable model. Read more - [Enerace case study](https://ordergroup.co/case-studies/enerace-energy-trading-software/) - [Centrum Komunikacji case study](https://ordergroup.co/case-studies/centrum-komunikacji-sign-language-interpreting-platform/) ![Four integrations in live and in-progress products](https://ordergroup.co/media/images/photo-1551288049-be.2e16d0ba.fill-2000x1200.format-webp.webp) ## What each integration looks like Since 2020 we have been building and developing [Enerace.Online](https://ordergroup.co/case-studies/enerace-energy-trading-software/), an advisory platform for industrial buyers of energy and gas. In its market analysis, news items are pinned to the price charts and each one gets an AI-generated summary the moment it is added. The feature has been live since 2025, in a product with more than 20 functional areas in one codebase, and it came without rewriting the platform. The stack is Python and Django with a GraphQL API, React with TypeScript, React Native, Celery and AWS. Centrum Komunikacji runs a Polish Sign Language (PJM) interpreting platform that we build in a government accessibility program (see the [case study](https://ordergroup.co/case-studies/centrum-komunikacji-sign-language-interpreting-platform/)). We added live speech-to-text for hard-of-hearing and late-deafened users. Automatic speech recognition writes a first draft in real time, a transcription specialist corrects it, and only the corrected text reaches the user, as a chat message. The feature runs in production. Personal data is protected by encrypted fields, the option to delete a user's data and an event log. On a product platform in the home renovation and interior fittings market, an AI chat and a product-search agent run in a separate AI service. In Generator [ETR](https://ordergroup.co/glossary/easy-to-read/) for PSONI the phone reads the text from a photo of a letter, on the device itself. PDF documents take a separate path on the server before they reach the model. The reading itself is done by ready-made system libraries, Google ML Kit and Apple Vision. The model connection, the text reading and the readability score are in place; the API for other systems is still being built. The AI part sits outside the application ## LLM integration: a separate AI service your app calls Your backend calls the AI service through its own service layer. The model sits behind it and can be changed without touching the app. Read more - [Private LLM deployment](https://ordergroup.co/ai-software-development/private-llm/) ![LLM integration: a separate AI service your app calls](https://ordergroup.co/media/images/llm_chat_desktop_1x_47EDKKL.2e16d0ba.fill-2000x1200.format-webp.webp) A separate AI service is our default architecture for LLM integration. The application's backend has a service layer that calls it, so we can measure the AI part, change it or switch it off without touching the rest of the product. On the renovation product platform, a flag at project level (ai_enabled) switches AI on, so the feature can be switched off without changing the rest of the product. ## Swapping the model without redeploying the app In Generator ETR for PSONI the connection to the model is a setting. An administrator changes the address, key, model name and generation settings such as temperature and max_tokens in the Django Admin panel, and the backend containers keep running. A model change therefore does not need a new release of the application. Which model sits behind the service is a separate choice. A model provider's API is faster to start. When your data cannot leave your infrastructure, we host the model ourselves; see [our page on private LLM deployment](https://ordergroup.co/ai-software-development/private-llm/). Either way, the AI part stays a separate service. An API for other systems and an MCP server ## Opening the AI feature to other systems PSONI's AI feature is getting an API for other systems. AI agents can reach a product through an MCP server. Read more - [MCP server development](https://ordergroup.co/ai-software-development/mcp-server-development/) - AI agent development ![Opening the AI feature to other systems](https://ordergroup.co/media/images/kaleidico-3V8xo5Gbu.2e16d0ba.fill-2000x1200.format-webp.webp) At PSONI's decision the API for other systems will have four endpoints: logging in with a token, generating an easy-to-read text, listing a user's texts and previewing a historical text, which is kept for 30 days. Access uses an API key from the Django Admin panel, and the endpoints can be switched off later. To connect AI agents to an existing product, we build an MCP server ([Model Context Protocol](https://ordergroup.co/glossary/model-context-protocol/), a standard that lets AI agents call a product's functions) as a thin layer over its REST API. In our own CMS every MCP tool calls an API endpoint, so validation, permissions and logging stay in the API. Agents and chat assistants have their own page, AI agent development, and the server side is covered in [MCP server development](https://ordergroup.co/ai-software-development/mcp-server-development/). ## Where a person checks the AI Where a wrong answer has consequences for users, a person or a test checks the output. At Centrum Komunikacji only the text that a transcription specialist has corrected reaches the user. In Generator ETR a script scores each text with the Gunning Fog index, a readability measure, and a reviewer of easy-to-read texts rates real outputs. Jailbreak tests, which try to make the model break its rules, run on every new model iteration. Have a product that already works and a task such as answering questions or transcribing speech? Start with a free consultation. [Book a free consultation](https://ordergroup.co/contact-us/) ## Integration options compared Five options compared by what each is and what to know before choosing it. | Option | What it is | What to know | |---|---|---| | Separate AI service | Your backend calls the AI part through its own service layer | Our default architecture. The AI part can be measured, changed or switched off on its own. | | Model provider's API | The AI service sends requests to the API of a model provider | Faster to start. Your data leaves your infrastructure. | | Model we host | We run the model ourselves instead of calling a provider's API | The choice when data cannot leave your infrastructure. See private LLM deployment. | | API for other systems | Endpoints that other systems call with a key from the Django Admin panel | At PSONI: four endpoints, texts kept for 30 days, endpoints that can be switched off later. | | MCP server | A thin layer over the product's REST API that AI agents use as tools | Validation, permissions and logging stay in the API. | ## How an integration project runs ### AI readiness: what we check before implementation A free consultation leads to a free [discovery workshop](https://ordergroup.co/services/consulting-discovery/product-workshops/), which is our AI readiness assessment. We look at four things. - The product: what it does today and where an AI feature would sit in it. - The data: what the model would read, and whether that data may leave your infrastructure. - The rules: the legal, security and company constraints your product works under. - The task and the test: what the AI should do, and which test will show that it works. Sometimes the outcome is advice to skip the language model altogether. You leave the workshop knowing what the AI is meant to do and how its result will be judged. ### AI implementation: stages 1. Discovery. A free workshop, run with our own AI toolkit, that fixes the scope and the success test. 2. A separate AI service. We build the AI part as a service your backend calls, so the rest of the product stays untouched. 3. Model choice. A model provider's API or a model we host, depending on where your data may go. 4. Quality checks. The test from discovery runs on the output, and where an error is costly a person reviews it as well. 5. Support. After launch we stay on for [support and maintenance](https://ordergroup.co/services/platform-cloud-engineering/support-maintenance/), because models, prompts and the data they read keep changing. The scope of AI work is set in discovery. That scope covers the AI service, the model choice and the checks on its output. Copyright to the code passes to you with each payment, within the scope of use defined in the contract. Fixed test, owned code, support after launch ## Why teams add AI with us The success test is fixed in a free discovery workshop, before the build starts. The code is yours with each payment, and we stay on for support after launch. Read more - [Free discovery workshop](https://ordergroup.co/services/consulting-discovery/product-workshops/) - [Support and maintenance](https://ordergroup.co/services/platform-cloud-engineering/support-maintenance/) - [AI software development](https://ordergroup.co/ai-software-development/) ![Why teams add AI with us](https://ordergroup.co/media/images/discovery_people_ta.2e16d0ba.fill-2000x1200.format-webp.webp) Software projects **160+** delivered since 2013 Industries **15+** where we have built software App users **40,000+** people using apps we built, across all our projects Clutch rating **5.0/5 on Clutch** average rating of our client reviews on Clutch Clutch reviews **18 verified reviews** verified client reviews Do you have questions? ## AI integration services: frequently asked questions ![Łukasz Gajownik](https://ordergroup.co/media/images/T02DHCC1Z-UT420UBA7-d74022d27c3c-512.format-webp.webp) Łukasz Gajownik Head of AI & Frontend [Book a free consultation](https://ordergroup.co/contact-us/) ### What do AI integration services include? They cover a free discovery workshop, a separate AI service that your application calls, a choice between a model provider's API and a model we host, checks on the output, and support after launch. They suit a product owner or CTO whose web or mobile product already works and who wants to add AI without breaking the rest. ### Can you add AI to our existing product without rebuilding it? Yes. Enerace gets an AI summary of each market news item, added without a rewrite to a platform with more than 20 functional areas in one codebase. Centrum Komunikacji got live speech-to-text on an existing platform, and a renovation product platform got an AI chat and a product-search agent. ### How do you integrate an LLM into an existing application? We build the LLM part as a separate AI service that your backend calls through its own service layer. The connection to the model (address, key and model name) can be set in the Django Admin panel, as it is in PSONI's Generator ETR. ### Can we switch the AI feature off? Yes. On the renovation product platform a flag at project level (ai_enabled) switches AI on, so turning it off changes nothing else in the product. The AI part is a separate service, so it can also be measured or changed on its own. ### Can we change the model later? Yes. In Generator ETR an administrator changes the model name, address, key and generation settings in the Django Admin panel, and the backend containers keep running. A model change does not need a new release of the application. ### Model provider's API or our own model? A model provider's API is faster to start. When your data cannot leave your infrastructure, we host the model ourselves. We decide this with you in discovery, based on where the data may go. [Our page on private LLM deployment](https://ordergroup.co/ai-software-development/private-llm/) covers the second option. ### How do you check AI output inside our product? Where a wrong answer has consequences for users, a person or a test checks the output. At Centrum Komunikacji a transcription specialist corrects the speech-to-text before the user sees it. In Generator ETR we score texts with the Gunning Fog index, a reviewer rates them, and jailbreak tests run on every new model iteration. ### Do you build RAG systems? Yes. We settle the scope and architecture of a RAG system in the discovery workshop. We do not describe RAG projects here because they are covered by NDA. The questions are the same as for any AI feature: what the system should do and which test will show that it works. ### What is an AI readiness assessment? It is the check we run in the free discovery workshop before any AI implementation. We look at your product, your data and the rules you work under, then agree what the AI should do and which test will show that it works. Sometimes the answer is to skip the language model. ### What are the stages of AI implementation? Discovery comes first and is free. Then we build a separate AI service, choose between a model provider's API and a model we host, check the output with a test and a person where an error is costly, and stay on for support after launch. ### Who owns the code of an AI integration? You do. Copyright to the code passes to you with each payment, within the scope of use defined in the contract. This applies to the code we write for the AI integration, in the AI service as well as in your backend. ### Which technologies do you use for AI integration? It depends on your product. In Enerace the stack is Python and Django with a GraphQL API, React with TypeScript, React Native, Celery and AWS. In Generator ETR the model settings live in Django Admin, and text is read from photos with Google ML Kit and Apple Vision on the phone. ![Developer with headphones coding at desktop monitor in an open office](https://ordergroup.co/media/images/Image_1_3kejLC8_Y69T2Sz.format-webp.width-2880.webp) ## AI features added to products that already work. Your message goes to ![Filip Szamborski](https://ordergroup.co/media/images/Filip_dTWHyBo.2e16d0ba.fill-432x432.format-webp.webp) Filip Szamborski Co-founder & CEO [](https://www.linkedin.com/in/filip-szamborski/) SO FAR WE HAVE WORKED WITH BRANDS LIKE: - [![RAW Cyber](https://ordergroup.co/media/images/raw-cyber-mono_M2DWvZf_pHEQtjP.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/raw-cyber-custom-android-os/) - [![Skyfri](https://ordergroup.co/media/images/skyfri-mono_KGicBgp.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/skyfri-solar-asset-management-platform/) - ![Spacestor](https://ordergroup.co/media/images/spacestor-mono_woFaKhf.max-208x64.format-webp.webp) - [![Zeronest](https://ordergroup.co/media/images/zeronest-mono_BqSPQJQ_2nGxsSc.max-208x64.format-webp.webp)](https://ordergroup.co/case-studies/zeronest-solar-panel-monitoring-app/) - ![AASA24](https://ordergroup.co/media/images/aasa24-mono.max-208x64.format-webp.webp) Thank you for your message. We read every message ourselves and get back to you within 48 hrs on business days. Something urgent? [hello@ordergroup.co](mailto:hello@ordergroup.co)