# AI, applied with judgement

> AI-centred products, agent-ready websites, and AI tools applied properly: three ways I help teams use AI where it genuinely pays its way.

I was shipping NLP and document-intelligence systems years before the boom. That history buys two things: I know what AI is genuinely good at, and I know what it quietly fails at.

## AI-centred products

Products where AI is the core of the value, not a bolt-on: agentic systems, LLM pipelines, retrieval and search over messy real-world data, with the engineering that makes them dependable (evaluation, guardrails, observability, cost control). Proof: a platform streaming millions of documents a month through NLP grading at S-RM. Case study: https://blizzard.consulting/work/document-intelligence-at-scale.md

## Agent-ready websites & applications

AI agents now read sites, answer questions from them, and act on them for their users. This site is the demonstration: llms.txt (https://blizzard.consulting/llms.txt), markdown for every core page via Accept: text/markdown, content signals in robots.txt, agent skills (https://blizzard.consulting/.well-known/agent-skills/index.json), and an API catalogue (https://blizzard.consulting/.well-known/api-catalog). I audit and retrofit existing sites the same way. Standards: https://specification.website/spec/agent-readiness/

## AI tools, applied properly

AI applied carefully inside the business you already run: agentic search across research libraries and document-heavy systems, AI-assisted engineering with a senior engineer accountable for every line, and an honest map of where AI does not belong yet.

## Contact

- Email: jim@blizzard.consulting
- Canonical page: https://blizzard.consulting/ai/