> Markdown version of [/videos/1513-best-practices-building-enterprise-applications-that-leverage-genai?t=868](https://www.wearedevelopers.com/videos/1513-best-practices-building-enterprise-applications-that-leverage-genai?t=868). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Best practices: Building Enterprise Applications that leverage GenAI

Ready to ditch rigid UIs for natural language? Learn how RAG, Semantic Kernel, and function calling securely power enterprise apps without expensive model retraining.

- **Speakers:** [Damir](https://www.wearedevelopers.com/@damir)
- **Event:** World Congress 2025
- **Published:** August 20, 2025
- **Duration:** 24:24
- **URL:** https://www.wearedevelopers.com/videos/1513-best-practices-building-enterprise-applications-that-leverage-genai

## Summary

The evolution of enterprise software is transitioning from traditional graphical user interfaces to natural language-driven experiences, often referred to as 'Software v2'. Developers can now leverage large language models (LLMs) to build symbiotic connections between human language and system actions. Using .NET and Microsoft's Semantic Kernel, applications can seamlessly execute backend tasks—translating non-English commands to trigger physical smart devices or querying complex PowerShell metrics—without relying on rigid UI controls.

Extending base models without undertaking a prohibitively expensive retraining process (which can cost upwards of $20 million) relies on two primary methodologies: retrieval-augmented generation (RAG) and function calling. Through RAG, enterprise architectures can split text into targeted token chunks and inject custom data into the LLM's prompt context, effectively overriding native model biases with proprietary company datasets. Concurrently, function calling bridges the gap between text intents and execution. This shifts traditional development paradigms; engineers must now meticulously document their classes and parameters, as the LLM routes user intents based directly on those code descriptions.

At the heart of these semantic architectures are embeddings, which map tokens into 1,536-dimensional coordinate systems to calculate similarity distances. Robust enterprise solutions require dedicated vector databases, such as Qdrant, Cosmos DB, or the native vector search arriving in SQL Server 2025. Developers must aggressively optimize these systems, as unoptimized vector searches can degrade system performance from mere seconds to minutes as database scale grows. Ultimately, data privacy concerns are mitigated because LLMs remain fundamentally stateless; sensitive data injected at runtime only briefly interacts with the model to complete a user prompt before being discarded.

**Keywords:** enterprise ai applications, semantic kernel, RAG implementation, llm function calling, vector databases, token embeddings, sql server vector search, text chunking strategies, stateless ai compliance, natural language interfaces, software evolution, ai code documentation, powershell automation, azure ai foundry, token similarity distance

## Chapters

1. **Building enterprise artificial intelligence software in .NET** (00:05) — Integrating generative AI capabilities into enterprise systems requires understanding specific application frameworks and platforms like C#.
1. **Demonstrating natural language interactions with physical hardware** (01:37) — Leveraging semantic kernel bridges human language prompts seamlessly to practical software and hardware executions such as smart lighting.
1. **Mapping natural language prompts to PowerShell commands** (05:54) — User intents can accurately trigger command line scripts and system queries using integrated language models and plugins.
1. **Understanding embeddings and semantic token descriptions** (08:34) — Transforming tokens into mathematically measurable vectors enables systems to accurately gauge the semantic similarity between different texts.
1. **Storing and searching data with vector databases** (10:31) — Expanding massive array stores requires specialized databases capable of executing native vector distance functions efficiently to avoid heavy latency.
1. **Extending large language models without expensive retraining** (12:50) — Capabilities of pre-trained models are safely expanded using knowledge tools like retrieval augmented generation and action tools like function calling.
1. **Implementing text chunking and retrieval augmented generation** (14:28) — Custom C# applications calculate token consumption and segment raw text sequentially to map exact vector matches for incoming semantic queries.
1. **Overriding language model biases using custom vectors** (18:17) — Injecting custom business information directly into databases safely overwrites outdated training bias without incurring expensive retraining cycles.
1. **Executing system commands through model function calling** (19:26) — External plugins interpret tool assignments from the logic model to safely query APIs while tracking when sensitive data pipelines are touched.
1. **Building and debugging Semantic Kernel plugins programmatically** (21:18) — Detailed parameter descriptions inside backend classes ensure models accurately map natural language queries directly into system dependencies.

## Related Moments

- [Embedding generative AI in enterprise software platforms](https://www.wearedevelopers.com/videos/916-beyond-the-hype-real-world-ai-strategies-panel) (from "Beyond the Hype: Real-World AI Strategies Panel")
- [Scaling generative AI use cases across large enterprises](https://www.wearedevelopers.com/videos/916-beyond-the-hype-real-world-ai-strategies-panel) (from "Beyond the Hype: Real-World AI Strategies Panel")
- [Expanding AI capabilities using retrieval-augmented generation](https://www.wearedevelopers.com/videos/950-supercharge-your-cloud-native-applications-with-generative-ai) (from "Supercharge your cloud-native applications with Generative AI")
- [Audience questions on data integration and future action models](https://www.wearedevelopers.com/videos/899-creating-industry-ready-solutions-with-llm-models) (from "Creating Industry ready solutions with LLM Models")
- [Balancing human-centric AI collaboration with environmental sustainability practices](https://www.wearedevelopers.com/videos/1016-insight-into-ai-driven-design) (from "Insight into AI-Driven Design")
- [Navigating the components of the modern generative AI stack](https://www.wearedevelopers.com/videos/1512-building-ai-applications-with-langchain-and-node-js) (from "Building AI Applications with LangChain and Node.js")

## Related Articles

- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [Graph and AI Trends 2026: Why Is AI Running but Not Yet Delivering?](https://www.wearedevelopers.com/magazine/680-graph-and-ai-trends-2026-why-is-ai-running-but-not-yet-delivering)
- [Exploring AI: Opportunities and Risks for Developers](https://www.wearedevelopers.com/magazine/522-exploring-ai-opportunities-and-risks-for-developers)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Principal Engineer - AI Search & Vector Infrastructure](https://www.wearedevelopers.com/jobs/ext/353953-principal-engineer-ai-search-vector-infrastructure) at **Redis**
- [Principal Engineer - AI Search & Vector Infrastructure](https://www.wearedevelopers.com/jobs/ext/319507-principal-engineer-ai-search-vector-infrastructure) at **Redis**
- [Principal Engineer - AI Search & Vector Infrastructure](https://www.wearedevelopers.com/jobs/ext/381484-principal-engineer-ai-search-vector-infrastructure) at **Redis**