Topic
- Contract review is a citation problem, not a summarisation problemA lawyer does not want a summary of the indemnity clause. They want to be taken to it, with the words intact, so they can form their own view. That distinction changes the whole build.2 min read· Legal· RAG· LLM workflows
- Document extraction at the loading bayLogistics paperwork is the canonical extraction problem, and the difficulty is never the model. It is that the input is a photograph taken in a lorry cab by someone in a hurry.2 min read· Logistics· Data and integration· LLM workflows
- Reconciliation: the AI use case nobody demosIt is unglamorous, it is everywhere in finance operations, and it is a better fit for a language model than most of what gets demonstrated on stage — provided the model never touches the arithmetic.2 min read· Finance· Data and integration· LLM workflows
- Model upgrades are breaking changesA new model version is a dependency upgrade to the least deterministic component in the system, and it usually arrives without a changelog that describes what will change about your behaviour.2 min read· LLM workflows· Delivery· AI in production
- The cost model of an LLM feature, worked throughTeams estimate token cost and are surprised by the bill. The gap is almost never the price per token — it is retries, context growth, and the runs nobody counted.2 min read· Cost· LLM workflows· Platform
- Prompts are code, and they need the same disciplineA prompt change alters system behaviour in production. If it can reach users without review, a test, or a way to roll back, then the most volatile part of your system is also the least governed.2 min read· LLM workflows· Delivery
- Structured output is the difference between a demo and a featureA model that returns prose is a conversation. A model that returns a validated object is a component you can build on. Getting from one to the other is mostly unglamorous contract work.2 min read· LLM workflows· Architecture
- RAG is a retrieval problem wearing a language model costumeWhen a retrieval-augmented system gives a bad answer, the model is almost never the reason. It answered correctly from the wrong documents, because retrieval handed it the wrong documents.2 min read· RAG· LLM workflows
- Where an LLM belongs in a workflow, and where it does notThe useful question is not whether a language model can do a step. It is whether the step tolerates being wrong occasionally, and what happens when it is.2 min read· LLM workflows· Architecture
