Insights

What’s newly possible. And what it changes.

Not AI news. Three kinds of writing from the engineers who build these systems:

The Frontier: What’s newly possible, and what it changes for a real operation.
Applied: How new capabilities become production systems.
In Practice: What companies are actually doing with it.

Lead story /Applied

RAG Pipelines in Legacy System Integrations

How RAG pipelines connect AI to legacy systems to cut latency, boost accuracy, and reduce costs while preserving compliance.

What’s newly possible, and what it changes for a real operation.

The first Frontier pieces are being written: one new capability each, and what it changes for a real operation.

How new capabilities become production systems.

What companies are actually doing with it.

Browse by capability

Agentic workflows
Multi-step processes run end to end. People approve the exceptions instead of doing every step.
- Semantic Kernel in RAG API Architectures
- Unveiling Semantic Kernel
Computer use
AI operates software that has no API, so a legacy system stops being the reason you can’t automate.
First piece in progress
Voice agents
The phone becomes a channel AI can answer, qualify, and act on while the caller is still on the line.
First piece in progress
MCP & agent interoperability
Your software gains a second kind of user: agents, with scoped and auditable access to act.
First piece in progress
Multimodal AI
Documents, images, and forms become structured data without anyone keying them in.
- Azure AI Document Processing with .NET
- Supercharge Your C Sharp Apps with AI-Powered Parsing Using SmartParser
Reasoning systems
Models work through policy, rules, and edge cases, and show how they reached the answer.
First piece in progress