# Notes from the workbench.

Source: https://customlabs.io/insights/
Updated: 2026-08-31

Insights

# Notes from the workbench.

Working notes on shipping AI in production: what breaks between demo and deployment, what's worth buying instead of building, and how we test the parts that matter.

Methodology August 12, 2026 · 9 min read

## What Actually Moved in the EU AI Act's 2026 Deadline Change

Regulation (EU) 2026/1744 pushed the AI Act's high-risk deadline to December 2027. What moved, what didn't, and why the old schedule still pays off.

[Security & Governance](https://customlabs.io/topics/ai-security-governance/)

[Read →](https://customlabs.io/insights/ai-act-high-risk-deadline-moved/)

Methodology July 24, 2026 · 10 min read

## Prompt Injection Is a Data Problem: A Threat Model You Can Ship Against

[Prompt injection](https://customlabs.io/glossary/prompt-injection/) can't be filtered away: the model can't reliably tell instructions from data. Here's the actual threat model and the controls that hold up.

[Security & Governance](https://customlabs.io/topics/ai-security-governance/)[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)

[Read →](https://customlabs.io/insights/prompt-injection-threat-model/)

Teardown July 19, 2026 · 8 min read

## Why Your AI Agent Works in the Demo and Stalls in Production

An agent that nails the demo stalls in production because reliability compounds across steps. Here's the math, the real failure modes, and how to ship anyway.

[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)

[Read →](https://customlabs.io/insights/ai-agents-stall-in-production/)

Benchmark July 11, 2026 · 9 min read

## What AI Actually Costs in Production, by Workload

A reproducible benchmark of cost per successful outcome across four common AI workloads, with every token assumption, price, and overhead multiplier shown.

[AI Cost & Efficiency](https://customlabs.io/topics/ai-cost-efficiency/)[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)

[Read →](https://customlabs.io/insights/ai-cost-benchmark-by-workload/)

Build / Buy / Skip July 10, 2026 · 7 min read

## Model-Agnostic by Design

Models change under you every few months: price, quality, and capability. Here's why we never hardcode a single provider into a client's feature.

[Strategy & Architecture](https://customlabs.io/topics/strategy-architecture/)

[Read →](https://customlabs.io/insights/model-agnostic-by-design/)

Teardown July 7, 2026 · 7 min read

## Your RAG Demo Lied to You

Retrieval that looks flawless on ten clean PDFs falls apart on a real corpus. Here's why, and what evaluating retrieval quality actually requires.

[Retrieval & RAG](https://customlabs.io/topics/retrieval-rag/)[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)

[Read →](https://customlabs.io/insights/your-rag-demo-lied/)

Benchmark June 30, 2026 · 8 min read

## What an AI Feature Actually Costs in Production

Token costs that look trivial in a demo compound fast at scale. Here's how to make cost a first-class metric instead of a surprise on the invoice.

[AI Cost & Efficiency](https://customlabs.io/topics/ai-cost-efficiency/)[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)[Evals & Observability](https://customlabs.io/topics/evals-observability/)

[Read →](https://customlabs.io/insights/what-ai-actually-costs/)

Methodology June 23, 2026 · 7 min read

## Evals Before You Ship: Why AI Features Need Tests Too

Shipping an AI feature without an [eval suite](https://customlabs.io/glossary/eval-suite/) in CI means every prompt tweak is a guess. Here's what an eval suite actually needs to cover.

[Evals & Observability](https://customlabs.io/topics/evals-observability/)

[Read →](https://customlabs.io/insights/evals-before-you-ship/)

Build / Buy / Skip May 19, 2026 · 7 min read

## Build, Buy, or Skip: A Framework for AI Decisions

A technically honest framework for deciding whether an AI initiative should be built in-house, bought off the shelf, or skipped entirely this year.

[Strategy & Architecture](https://customlabs.io/topics/strategy-architecture/)[AI Cost & Efficiency](https://customlabs.io/topics/ai-cost-efficiency/)

[Read →](https://customlabs.io/insights/build-buy-or-skip/)

Teardown April 14, 2026 · 7 min read

## From Notebook to Production: What Actually Breaks

The gap between a working AI demo and a production feature is auth, latency, cost and fallbacks. Here's how we close it without a rewrite.

[Shipping to Production](https://customlabs.io/topics/shipping-to-production/)

[Read →](https://customlabs.io/insights/notebook-to-production/)

No insights match that filter yet.

Newsletter

Notes from the workbench.

One email when we publish something worth reading. No drip campaign.

Website

Subscribe →