Guides for putting AI to work
Plain-English guides for teams adopting AI, and engineering notes from the systems we build: AI agents, support and voice automation, production RAG, streaming, evaluation and MCP.
Kompactor: Cut the Tokens Claude Code Spends on Noise
How Claude Code spends tokens on tool output, how the free Kompactor plugin shortens it with hooks, and what 54 measured sessions showed: −16% tokens, −10% cost, same answers.
Read the guideGemini 4 Argon vs Opus 5.5 vs Fable 5.1 vs GPT-6 Astra
Gemini 4 Argon benchmarks vs Claude Opus 5.5, Fable 5.1 and GPT-6 Astra: coding, long context, hallucinations, pricing, and the future of coding.
Read the guideGDPR-Compliant AI Agents and Chatbots: A UK and EU Guide
How to build GDPR-compliant AI chatbots and AI agents in the UK and EU: data residency, DPAs, retention, EU AI Act transparency, and a practical checklist.
Read the guideAI Customer Support Agents: Resolve Half Your Tickets Without Hurting CSAT
How AI customer service agents work, what they can resolve, helpdesk AI vs custom agents, realistic resolution rates, costs, and a rollout plan that protects CSAT.
Read the guideAI Receptionist for Small Business: Cost, Setup, and ROI
Should your small business use an AI receptionist? Costs vs a human receptionist or answering service, what it can handle, setup steps, and an ROI example.
Read the guideAI Voice Agents for Business: How They Work, Costs, and Platforms
A practical guide to AI voice agents and AI phone agents: how they work, real use cases, per-minute costs, Vapi vs Retell vs LiveKit, and how to launch one.
Read the guideAI Agents for Business: Use Cases, Costs, and How to Start
What AI agents for business actually do in 2026: 10 real use cases, how they work, what they cost to build and run, and a 30-day plan to launch your first one.
Read the guidePrompt Engineering vs AI Agents
Prompt engineering is one-shot instruction. AI agents loop with tools. You still write prompts inside agents — evals and tool design beat magic templates.
Read the guideWhat Is RAG
RAG is retrieval-augmented generation: search your docs, then ask an LLM. Pipeline, embeddings, chunking, failure modes, and a tiny Python example.
Read the guideAI Agent Testing and Evaluation Strategies
Production AI agent testing — golden dataset eval, unit tests for tools, trace replay, regression gates, and continuous monitoring patterns for reliable agent deployment.
Read the guideMulti-Agent Pipeline for SEO and Content Automation
Multi-agent pipeline for SEO and content automation — five agent roles, metadata checklist, static site workflow, and GSC feedback loop with copy-paste patterns.
Read the guideHow to Build a Cursor-Like AI Coding Agent — Complete Guide
Step-by-step guide to building a Cursor-like AI coding agent: tool-calling harness loop, SSE streaming, context compaction, permissions, subagents — with full pseudocode from LiveCode.
Read the guideHow I Built a Live-Browser AI Coding Agent with Flask and Monaco
Build a Cursor-style AI coding agent in the browser: Flask harness loop, SSE + WebSocket streaming, Monaco editor, terminal, JSONL sessions, and three-tier context compaction.
Read the guideGPT-5.6 Sol vs Claude Fable 5 vs Kimi K3: Which Should You Use?
Compare GPT-5.6 Sol, Claude Fable 5, and Kimi K3 — coding agents, reasoning, pricing, open weights, and which model to use for each job in 2026.
Read the guideAuto Model Routing Without Calling an LLM to Pick an LLM
Route requests between cheap and expensive models using heuristics instead of a classifier call — rules, thresholds, and the cost maths behind each hop.
Read the guideSSE vs WebSocket vs REST API for Live AI
A decision table for SSE, WebSockets and REST on live AI workloads — latency, reconnection, proxy behaviour, FastAPI code for each, and when a hybrid wins.
Read the guideAutomate Data Engineering Fixes with AI Agents and MCP
Automate data engineering fixes with AI agents and MCP — Jira, Confluence, GitHub CI, SonarQube, Airflow, and MongoDB in a retry loop until DAGs are green.
Read the guideHow SSE Streaming Works in Chatbots
Learn how SSE streaming works in chatbots — real LLM token streams over Server-Sent Events, FastAPI API setup, and wiring OpenAI, Claude, or Gemini.
Read the guideBest Economical LLM Models for RAG
Best economical LLM models for RAG in 2026 — compare Gemini, GPT, and Claude API pricing with tables, charts, and cost per turn with chat history.
Read the guideCohere Reranking for Production RAG
Cohere reranking for production RAG retrieval — two-stage vector search, rerank-v3.5, conditional skip, MMR diversity, query rewrite, and cost-aware gates.
Read the guideHow to Evaluate RAG Retrieval
How to evaluate RAG retrieval before tuning prompts — golden eval sets, hit@k, MRR, failure taxonomy, and a weekly loop for measurable search quality.
Read the guideBuild a RAG Chatbot in Python: Complete Production Guide
Build a RAG chatbot in Python step by step — chunking, embeddings, vector database, retrieval, LLM answers, evaluation and production deployment, with full code.
Read the guideNo posts match your search.
Planning an AI project?
Book a free 30-minute intro call. Bring one process you'd like AI to take off your team's plate, and leave with a clear next step.