Blog

Our writing on AI — implementation, strategy, and the messy in-between

Practical posts on building and deploying AI. No predictions about AGI. Lots of detail on the parts that actually trip companies up.

AI Strategy Posts

Posts on strategy. Practical, opinionated, written by the people doing the work.

Explore

AI Implementation Posts

Posts on implementation. Practical, opinionated, written by the people doing the work.

Explore

Project Write-Ups

Posts on case-studies. Practical, opinionated, written by the people doing the work.

Explore

AI Trends & Commentary

Posts on trends. Practical, opinionated, written by the people doing the work.

Explore

AI Engineering Deep Dives

Posts on engineering. Practical, opinionated, written by the people doing the work.

Explore

Why Most AI Pilots Die in Procurement

Why Most AI Pilots Die in Procurement — a practical post from our team, drawn from real engagements.

Explore

Five RAG Architecture Mistakes We Keep Seeing

Five RAG Architecture Mistakes We Keep Seeing — a practical post from our team, drawn from real engagements.

Explore

Should You Fine-Tune Claude or Just Prompt It?

Should You Fine-Tune Claude or Just Prompt It? — a practical post from our team, drawn from real engagements.

Explore

What the EU AI Act Actually Means for Mid-Market Companies

What the EU AI Act Actually Means for Mid-Market Companies — a practical post from our team, drawn from real engagements.

Explore

The Real Cost of an AI Agent in Production

The Real Cost of an AI Agent in Production — a practical post from our team, drawn from real engagements.

Explore

How We Actually Evaluate LLM Applications

How We Actually Evaluate LLM Applications — a practical post from our team, drawn from real engagements.

Explore

Build vs Buy: Conversational AI

Build vs Buy: Conversational AI — a practical post from our team, drawn from real engagements.

Explore

Your Data Warehouse Is Not RAG-Ready

Your Data Warehouse Is Not RAG-Ready — a practical post from our team, drawn from real engagements.

Explore

The Quietest LLM Cost Blowups (And How to Spot Them)

The Quietest LLM Cost Blowups (And How to Spot Them) — a practical post from our team, drawn from real engagements.

Explore

When Not to Use an LLM

When Not to Use an LLM — a practical post from our team, drawn from real engagements.

Explore

AI Chatbot Anti-Patterns We Won't Ship

AI Chatbot Anti-Patterns We Won't Ship — a practical post from our team, drawn from real engagements.

Explore

A Practitioner's Vector DB Shootout

A Practitioner's Vector DB Shootout — a practical post from our team, drawn from real engagements.

Explore

How an Eval Set Changed Everything for One Project

How an Eval Set Changed Everything for One Project — a practical post from our team, drawn from real engagements.

Explore

Five Things That Move Retrieval Quality the Most

Five Things That Move Retrieval Quality the Most — a practical post from our team, drawn from real engagements.

Explore

Agent Loops That Don't Converge (and What We Do About It)

Agent Loops That Don't Converge (and What We Do About It) — a practical post from our team, drawn from real engagements.

Explore

Classical ML Isn't Dead — Three Recent Wins

Classical ML Isn't Dead — Three Recent Wins — a practical post from our team, drawn from real engagements.

Explore

AI Procurement: Questions Vendors Hate

AI Procurement: Questions Vendors Hate — a practical post from our team, drawn from real engagements.

Explore

The Model Router Is a Real Pattern Now

The Model Router Is a Real Pattern Now — a practical post from our team, drawn from real engagements.

Explore

When Fine-Tuning Beats RAG (and Vice Versa)

When Fine-Tuning Beats RAG (and Vice Versa) — a practical post from our team, drawn from real engagements.

Explore

How We Staff an AI Engagement

How We Staff an AI Engagement — a practical post from our team, drawn from real engagements.

Explore

Our Discovery Sprint Format, Step by Step

Our Discovery Sprint Format, Step by Step — a practical post from our team, drawn from real engagements.

Explore

The AI Team of One Playbook

The AI Team of One Playbook — a practical post from our team, drawn from real engagements.

Explore

Hallucination Isn't Just a Prompt Problem

Hallucination Isn't Just a Prompt Problem — a practical post from our team, drawn from real engagements.

Explore

Voice Agents and the Latency Budget Problem

Voice Agents and the Latency Budget Problem — a practical post from our team, drawn from real engagements.

Explore

Multi-Agent Systems vs Pipelines, Honestly

Multi-Agent Systems vs Pipelines, Honestly — a practical post from our team, drawn from real engagements.

Explore

Evaluating RAG the Grown-Up Way

Evaluating RAG the Grown-Up Way — a practical post from our team, drawn from real engagements.

Explore

Structured Output: Best Practices Across Models

Structured Output: Best Practices Across Models — a practical post from our team, drawn from real engagements.

Explore

Internal Copilot Rollouts: The Boring Stuff That Matters

Internal Copilot Rollouts: The Boring Stuff That Matters — a practical post from our team, drawn from real engagements.

Explore

AI Cost Attribution Models for FinOps Teams

AI Cost Attribution Models for FinOps Teams — a practical post from our team, drawn from real engagements.

Explore

The LLM Router + Fallback Pattern

The LLM Router + Fallback Pattern — a practical post from our team, drawn from real engagements.

Explore

RAG Corpus Hygiene: The Job Nobody Names

RAG Corpus Hygiene: The Job Nobody Names — a practical post from our team, drawn from real engagements.

Explore

When to Add a Reranker (and Which One)

When to Add a Reranker (and Which One) — a practical post from our team, drawn from real engagements.

Explore

Synthetic Data, Honestly

Synthetic Data, Honestly — a practical post from our team, drawn from real engagements.

Explore

Internal AI Coding Assistants Worth Building

Internal AI Coding Assistants Worth Building — a practical post from our team, drawn from real engagements.

Explore

Feature Flags for AI: A Necessary Habit

Feature Flags for AI: A Necessary Habit — a practical post from our team, drawn from real engagements.

Explore

How to Write an Eval That Doesn't Lie to You

How to Write an Eval That Doesn't Lie to You — a practical post from our team, drawn from real engagements.

Explore

The Prompt as a Specification

The Prompt as a Specification — a practical post from our team, drawn from real engagements.

Explore

Honest Metrics for AI Customer Service

Honest Metrics for AI Customer Service — a practical post from our team, drawn from real engagements.

Explore

RAG vs Search, Honestly

RAG vs Search, Honestly — a practical post from our team, drawn from real engagements.

Explore

The 'Per-Seat AI' Pricing Trap

The 'Per-Seat AI' Pricing Trap — a practical post from our team, drawn from real engagements.

Explore

Self-Host or SaaS Your LLM? A Decision Tree

Self-Host or SaaS Your LLM? A Decision Tree — a practical post from our team, drawn from real engagements.

Explore

The Quiet Rise of Small Models in Production

The Quiet Rise of Small Models in Production — a practical post from our team, drawn from real engagements.

Explore

AI Document Extraction Tips We Wish We'd Known

AI Document Extraction Tips We Wish We'd Known — a practical post from our team, drawn from real engagements.

Explore

IDP vs Generic LLM Extraction

IDP vs Generic LLM Extraction — a practical post from our team, drawn from real engagements.

Explore

AI Content Pipeline Architecture for Marketing Teams

AI Content Pipeline Architecture for Marketing Teams — a practical post from our team, drawn from real engagements.

Explore

Voice of Customer Without the Buzzwords

Voice of Customer Without the Buzzwords — a practical post from our team, drawn from real engagements.

Explore

Brand-Safe LLM Content at Scale

Brand-Safe LLM Content at Scale — a practical post from our team, drawn from real engagements.

Explore

AI Translation Can't Replace Localization (Yet)

AI Translation Can't Replace Localization (Yet) — a practical post from our team, drawn from real engagements.

Explore

Evals for Agents vs Chatbots: They're Different

Evals for Agents vs Chatbots: They're Different — a practical post from our team, drawn from real engagements.

Explore

The Real Work of Data Prep for AI Projects

The Real Work of Data Prep for AI Projects — a practical post from our team, drawn from real engagements.

Explore

On-Prem AI Still Makes Sense for These Sectors

On-Prem AI Still Makes Sense for These Sectors — a practical post from our team, drawn from real engagements.

Explore

Model Cards as a Forcing Function

Model Cards as a Forcing Function — a practical post from our team, drawn from real engagements.

Explore

Postmortems for AI Incidents

Postmortems for AI Incidents — a practical post from our team, drawn from real engagements.

Explore

Prompt Injection Defenses That Actually Work

Prompt Injection Defenses That Actually Work — a practical post from our team, drawn from real engagements.

Explore

AI Data Leakage Patterns We've Seen

AI Data Leakage Patterns We've Seen — a practical post from our team, drawn from real engagements.

Explore

RAG and Permissions: A Solvable Problem

RAG and Permissions: A Solvable Problem — a practical post from our team, drawn from real engagements.

Explore

The AI Monitoring Stack: Real Options Today

The AI Monitoring Stack: Real Options Today — a practical post from our team, drawn from real engagements.

Explore

Notes from a Real OpenAI-to-Claude Migration

Notes from a Real OpenAI-to-Claude Migration — a practical post from our team, drawn from real engagements.

Explore

When to (Not) Write Your Own Vector DB

When to (Not) Write Your Own Vector DB — a practical post from our team, drawn from real engagements.

Explore

Dataset Versioning Isn't Optional

Dataset Versioning Isn't Optional — a practical post from our team, drawn from real engagements.

Explore

AI Agent Failure Modes, Categorised

AI Agent Failure Modes, Categorised — a practical post from our team, drawn from real engagements.

Explore

Evals as Tests in CI: A Working Pattern

Evals as Tests in CI: A Working Pattern — a practical post from our team, drawn from real engagements.

Explore

Scope Creep in AI Projects: Pre-Mortem

Scope Creep in AI Projects: Pre-Mortem — a practical post from our team, drawn from real engagements.

Explore

When to Use Structured Extraction Over Free-Form Generation

When to Use Structured Extraction Over Free-Form Generation — a practical post from our team, drawn from real engagements.

Explore

How to Write a Good AI Discovery Doc

How to Write a Good AI Discovery Doc — a practical post from our team, drawn from real engagements.

Explore

The AI Talent Market in 2026, From the Inside

The AI Talent Market in 2026, From the Inside — a practical post from our team, drawn from real engagements.

Explore

AI Vendor Due Diligence Checklist We Use

AI Vendor Due Diligence Checklist We Use — a practical post from our team, drawn from real engagements.

Explore

The 'Second Pilot' Problem in AI Programs

The 'Second Pilot' Problem in AI Programs — a practical post from our team, drawn from real engagements.

Explore

When the Board Asks for an AI Roadmap

When the Board Asks for an AI Roadmap — a practical post from our team, drawn from real engagements.

Explore

How to Explain LLM Cost to a CFO

How to Explain LLM Cost to a CFO — a practical post from our team, drawn from real engagements.

Explore

AI Readiness Isn't a Spreadsheet

AI Readiness Isn't a Spreadsheet — a practical post from our team, drawn from real engagements.

Explore

AI CoE vs Federated: Pick Based on Culture

AI CoE vs Federated: Pick Based on Culture — a practical post from our team, drawn from real engagements.

Explore

Inside the Pilot Graveyard

Inside the Pilot Graveyard — a practical post from our team, drawn from real engagements.

Explore

AI Program vs Projects: A Different Operating Model

AI Program vs Projects: A Different Operating Model — a practical post from our team, drawn from real engagements.

Explore

How to Pick the First AI Use Case

How to Pick the First AI Use Case — a practical post from our team, drawn from real engagements.

Explore

The Myth of the Prompt Engineering Job

The Myth of the Prompt Engineering Job — a practical post from our team, drawn from real engagements.

Explore

Ready to ship AI, not slides?

Senior-only delivery. Fixed-scope pilots. Your data stays yours.

Subscribe to the newsletter