Tagged “decisions”
-
RAG or a Knowledge Graph?
Graph retrieval answers questions vector search structurally cannot. It also adds an extraction pipeline and a schema you own forever. When it pays.
-
Is RAG Still Worth Building?
Long context, agentic search and better models keep being called RAG's replacement. What actually goes away, what doesn't, and how to decide now.
-
When to Bring in Outside Help for Retrieval
Consultants can build a retrieval system faster than you can. Which parts to buy, which never to outsource, and the handover clause that decides it.
-
RAG Versus Fine-Tuning as an Architecture Decision
Not which is more accurate — which one leaves you able to fix a wrong answer on a Tuesday afternoon. An ownership-first comparison.
-
When Not to Use RAG
Four situations where retrieval is the wrong shape for the problem, what to build instead, and the threshold at which each answer flips.
-
Do You Need Generation, or Just Better Search?
Half the projects that ask for a RAG chatbot would be better served by a search page with good snippets. How to tell which one you have.
-
Build or Buy Your Retrieval Stack
What a managed retrieval product actually removes from your plate, what it never removes, and the four questions that decide the call.
-
When to Retire a RAG System
The decision nobody plans for. Five signals that a retrieval system has stopped earning its maintenance, and how to shut it down without losing the value.
-
Data Residency as an Architecture Constraint
When documents cannot leave a jurisdiction or a network, the retrieval decision changes shape. What the constraint actually forbids, and what it costs.
-
Scoping a Proof of Concept That Decides Something
Most RAG pilots prove only that the technology exists. How to scope one that answers the question you actually have to answer.
-
The Cost of Being Wrong
Put a number on what one bad answer costs you, and the architecture decision makes itself. A method for sizing error cost before you build.
-
One Index or Many?
Splitting a retrieval index by tenant or topic buys isolation and costs overhead. The three reasons that justify a split, and the ones that don't.
-
Do You Need a Vector Database at All?
A dedicated vector store is one of four places embeddings can live. Most teams are below the size where it wins, and the wrong choice adds an operator.
-
Pricing a RAG Project Before You Have Requirements
How to give a sponsor a defensible number when nobody has decided the scope yet — and which three questions collapse most of the uncertainty.
-
The Assistant You Already Pay For
Before you build retrieval, check what your existing software suite already does with your documents. The third option in most build-or-buy debates.
-
RAG or Tool Calls: Which One Does Your Question Need?
Retrieval reads documents; tool calls query systems. Most projects need both, and the split between them decides how much you have to maintain.
-
Choosing an Architecture by How It Fails
Every option for answering questions from your documents fails differently. Pick the failure mode your users and your business can actually survive.
-
The Team You Need to Run Retrieval in Production
Retrieval is not a project you staff once. Five capabilities the system needs permanently, and what to do when you can only fund three of them.