The Search Box Is the Most Expensive Feature Nobody Asked For
Two years ago, an operations manager at a logistics customer asked me for a feature, and I remember feeling relieved, because it sounded small. She said: I do not need another report. I need to type a phone number into one box and see everything that has ever touched that customer. Then she pointed at our product, which by then had thirty-four list views spread across eleven applications, each…
A logistics customer asked for a simple search box feature, but it proved to be surprisingly complex and costly. The request came from an operations manager who wanted to type a phone number into a box and see all related data. The existing platform had 34 list views across 11 applications, each with its own filter panel, but customers had to know the table a thing lived in to search for it.
The new global search box took 8 months to ship and ended up being the most expensive feature built that year. A filter is a structured question, while a search box is an unstructured question. Filters have known inputs and are easy to implement, while searches have to deal with unknowns. Search queries involve three different corpora: records, metadata, and unstructured content.
Each has its own tokenizer, permission rules, retention policies, and cost. Permissions are a major issue, as search indexes copy data out of the system of record, bypassing permission models. This creates leaks and potential security problems. Search also presents language and tokenization challenges, with different rules for various languages and types of data.
Freshness, backfill, and tenant onboarding add further complexity, making search a load-bearing component for AI agents.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
