Straterai/Field notes
Index · Field notes

Our learnings from the field.

Filter · Category
§ 05.01MAY · 15 · 20267 MIN
Primer·Foundations

How AI agents actually work: a non-technical primer.

'AI agent' is one of the most-used phrases in AI right now, and one of the most loosely defined. Here is the clearest one-sentence version, and how agents actually work underneath it.

By Charlie Stoner·AgentsTool-usePrimer
§ 05.02MAY · 15 · 20267 MIN
Primer·Foundations

What is RAG, and when does a business need it.

RAG, retrieval-augmented generation, is the most useful AI architecture for a business that has its own data and wants AI to actually use it. Here is what it is, how it works, and the three questions that decide whether your business needs one.

By James Finnegan·RAGWorking brainsArchitecture
§ 05.03MAY · 15 · 20265 MIN
Concept·AI for Business

What is an AI business operating system, and why companies are building them.

An AI business operating system is the connected layer a company's operations run on. Not a tool you add. The system underneath. Here is what one is, why a business reaches the point of needing one, and how a real one is built.

By James Finnegan·Operating systemsWorking brainsStrategy
§ 05.04MAY · 15 · 20265 MIN
Concept·AI for Business

What's the difference between an AI tool and an AI system.

An AI tool is a single function: a chatbot, a summarize button, a search box. An AI system is the architecture that connects functions together so they read your data, reason about it, and produce something an operator acts on.

By James Finnegan·Tools vs systemsStrategyPilots
§ 05.05MAY · 15 · 20264 MIN
Inside·Inside Straterai

We built a working brain for ourselves before we built one for anyone else.

We have a working brain. We built it months before we offered to build one for anyone else. We run the entire company on it.

By Charlie Stoner·Working brainFoundations
§ 05.06MAY · 15 · 20264 MIN
Architecture·Technical Insight

The model is the cheap layer.

When clients pay for production AI, they aren't paying for the model. They are paying for the architecture that catches what the model can't catch on its own.

By James Finnegan·HarnessArchitectureVendor evaluation
§ 05.07MAY · 15 · 20263 MIN
Architecture·Technical Insight

Tenant isolation belongs in the database, not the application.

Roughly seventy percent of multi-tenant AI applications enforce tenant isolation in application code. Every one of them is one missing filter away from a cross-tenant leak.

By Charlie Stoner·Tenant isolationSecurityArchitecture
§ 05.08APR · 18 · 20266 MIN
Architecture·Technical Insight

Why your AI tools are still answering, not deciding.

A retrieval app and an autonomous system look the same at first glance. They behave nothing alike on day 90.

By Charlie Stoner·AgentsArchitectureEnterprise
§ 05.09MAR · 30 · 20268 MIN
Operations·Technical Insight

A Postgres lake beats your data warehouse, for the things that matter.

Most enterprise data problems are operational, not analytical. Postgres handles them. Add the warehouse only when a specific workload proves it has to.

By James Finnegan·DataPostgres
§ 05.10MAR · 12 · 20265 MIN
Delivery·Inside Straterai

How enterprise AI systems actually ship in weeks, not months.

Every AI consulting deck right now claims weeks-not-months delivery. Most don't deliver. Three constraints separate the teams that do from the teams that don't.

By Charlie Stoner·DeliveryPatterns
§ 05.11FEB · 21 · 20265 MIN
Position·AI for Business

AI that understands your business is a system, not a chatbot.

Half the AI products that launched chat-first in 2024 are quietly shipping non-chat UI in 2026. The framing matters more than it sounds.

By James Finnegan·PositionStrategy
End · § 0511 / 11

Straterai Field Notes

Plain-English writing on building AI-native systems — how agents actually work, where they fail, and what we learn shipping them for real companies.

No spam. A couple of emails a month. Unsubscribe anytime.