Source: [AI that helps your team move the work forward.](https://treenodes.com/applied-ai/) This Markdown version is generated from the public page. Visit the source for interactive examples and full-size visuals. APPLIED AI # AI that helps your team move the work forward. Recommend useful content, turn documents into usable records and bring evidence together for decisions. TreeNodes designs, builds and supports AI within your software and workflows, with the permissions, human review and recovery each task needs. [Discuss your AI use case](https://treenodes.com/contact/?start=applied-ai)[Explore our AI services](https://treenodes.com/applied-ai/#capabilities-title) Diagram: Known inputs pass through a defined AI task, with human review where needed, producing a traceable result · Documents · Images · Records · Defined AI task · Human reviewwhen needed · Known inputs · Traceable result Known inputs → defined AI task → human review where needed → traceable result BUSINESS REQUIREMENTS & ANALYSIS ## Start with the work your business needs to do. We analyse how people work, the data they use and the business rules involved to define where AI can help. Together, we agree useful outcomes and the decisions that need human review. ### Recommend content and learning paths Recommend relevant content and learning paths using the information and preferences available in your application. ### Read documents and images Extract fields from invoices, scans and records. Return structured data with the source and flag what needs checking. ### Triage incoming work Classify requests, identify missing information and prepare the next step for the right person or workflow. ### Prepare evidence for decisions Bring relevant records together into a review, comparison or report that keeps supporting sources and unresolved questions visible. ### Act inside your systems Use permitted tools and services to carry out tasks, record results and hand over decisions that need human authority. ### Support the complete application Support the frontend, backend and data around AI, and implement new requirements as your business evolves. PROJECT SHOWCASE · MUSIC LEARNING Completed · 2023–early 2026 ## Personalised learning paths for musicians. TreeNodes delivered a complete music learning platform across web, iOS, iPad and Android, connecting course information, data processing and visualisation with personalised learning paths and certification. ![Music learning app views showing learning paths, course recommendations and learner data](https://treenodes.com/social/project-music-learning.png?v=fa7be7716536) Portfolio presentation of the delivered learning experience, with sample data and changed identities. ### The learning experience The platform combined an educational institution’s course information from databases, spreadsheets and other sources with student scores, interests, behaviour and learner characteristics. AI recommended courses and learning paths that musicians could follow towards certification. ### Web and native delivery The Swift apps for iOS and iPad and Kotlin apps for Android were delivered alongside the web platform by early 2026. The wider platform supported relationships across course, curriculum and learner information. Music learning platform **Inputs** Course and learning informationDatabases, spreadsheets and other sources from an educational institution **AI** Course and learning path recommendationsStudent scores, interests, behaviour and learner characteristics **Learning** Learning and certificationMusicians follow personalised learning paths, explore courses and gain certification Functional overview of the delivered platform. Client, platform and institution details remain confidential under NDA. **TreeNodes’ role** Analysis, architecture, frontend, backend, APIs & databases **Technology** C# · GraphQL · Swift for iOS and iPad · Kotlin for Android · Open-source AI API **Applied AI** Course and learning path recommendations [Explore the full project](https://treenodes.com/case-studies/music-learning/) Delivered under NDA. Client, platform and institution identities are not disclosed. IMPLEMENTATION, DELIVERY & OPERATION ## Carry the design through to working software. We build the user experience, backend, data connections and controls around AI, then plan release and ongoing support. **Build the complete workflow** Connect inputs, outputs, permissions and human review. Implement error handling and recovery for incomplete results or unavailable services. **Evaluate before release** Check quality, response time and cost against representative and difficult cases. Review errors with users and agree release criteria. **Operate and improve** Agree ownership and support, monitor performance and keep consequential actions traceable. Turn new requirements into controlled changes. ARCHITECTURE & API INTEGRATION ## Connect AI models to your business software. We integrate OpenAI, Claude and open-source AI models with your software, defining the context, data access, hosting and review each task needs. 1. 01 · Request ### Your website or app A person starts a task in your existing software. 2. 02 · Prepare ### Application backend Authentication Business rules Relevant context **Approved data** Documents or databases 3. 03 · Model request ### Choose a model API [**OpenAI** Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses) or [**Claude** Messages API](https://platform.claude.com/docs/en/api/overview) or [**Open-source AI** Hosted or self-hosted API](https://docs.vllm.ai/en/latest/serving/online_serving/openai_compatible_server/) 4. 04 · Return ### Backend validation The response returns to your backend for checks against the task and business rules. 5. 05 · Outcome ### Application result Show the checked result, or route it to a person for review. Human review when needed Conceptual integration: your backend controls access to approved data, sends the request to the chosen model API and checks the response before it reaches the application or a human reviewer. ### OpenAI Responses API · Model example: [GPT-6 Astra](https://developers.openai.com/api/docs/models) Reasoning, document analysis and tool use within your application and workflows. [OpenAI API documentation](https://developers.openai.com/api/docs/guides/migrate-to-responses) ### Claude Messages API · Model example: [Claude Opus 5](https://platform.claude.com/docs/en/models/overview) Document review, conversational applications and workflows that use tools. [Claude API documentation](https://platform.claude.com/docs/en/api/overview) ### Open-source AI Hosted or self-hosted model APIs Hosted APIs or models running in your infrastructure, selected for the task, licence terms, data requirements and running costs. [vLLM model-serving documentation](https://docs.vllm.ai/en/latest/serving/online_serving/openai_compatible_server/) We design credential handling, data access, structured responses, tool permissions, error recovery and usage monitoring around your application. Examples for new integrations, checked against official documentation on 13 September 2026. We select the model and configuration for each project. [Discuss your AI use case](https://treenodes.com/contact/?start=applied-ai)