Enterprise AI search can't answer the only question that matters: "why?"
Every company search tool demos the same way. Someone types "what's our refund policy?" , the AI returns a tidy paragraph with three source chips, the room nods, and procurement starts a twelve-week evaluation. Now watch an actual engineer's Tuesday: "Why is checkout rate-limited to 20 requests per second?" That answer is not in a document. It's in an ADR nobody linked, a Slack thread from…
Enterprise AI search tools claim to provide answers to company-wide queries, but they fall short when it comes to engineering questions that require understanding the "why" behind a system. These tools excel at retrieving relevant documents and summarizing information, but they struggle with tasks such as explaining design decisions, summarizing code, determining potential impacts of changes, and identifying the creators and owners of documents.
The issue lies in the fact that these tools primarily focus on retrieval and ranking of documents, rather than addressing the underlying problems that engineers are trying to solve. The article argues that enterprise search tools should adopt a graph-based approach, along with provenance tracking, to better handle the complex questions posed by engineers.
Additionally, the author suggests that these tools should have mechanisms to identify when information is missing or unverified, rather than presenting incorrect or hallucinated answers. The piece also touches on the pricing models of these tools, which often incentivize the tools to provide answers even when they lack sufficient information.
The author suggests that a bring-your-own-key approach, where users directly pay for the AI models they use, would encourage more responsible and transparent responses from the tools.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
