Finding: noise before signal.
General-purpose search indexes everything and has to guess at quality. So quality control happens late, by hand, and over and over again. Every project starts as a landfill to be sorted.
01 · The Problem
First, you throw away the web.
Ask the usual search box a research question and it answers from everywhere: adjacent disciplines, predatory journals, work that would never survive review. For a business scholar, the first job of every project is discarding what should never have appeared.
Fig. 1: a first page of results, elsewhere †
† Illustrative. Amber: the five results that belong. In Yajur, the rest never enter the room.
Reviewing: a method with no instrument.
A real systematic review is a procedure: protocol, search, screening, PRISMA, extraction, synthesis. Today it lives across reference managers, spreadsheets, and folders of PDFs, and routinely takes the better part of a year.
Comparing: forty PDFs, one table, by hand.
Theory, method, sample, findings: copied cell by cell into a spreadsheet nobody can audit. Tedious to build, impossible to check, and redone from scratch at every revision.
0→0
records screened → papers included, in a typical funnel
0+ months
a common timeline for one rigorous review
0
curated top-tier journals, a boundary nobody else indexes cleanly
These users judge a tool on whether it is correct and checkable. A tool that is confidently wrong even once loses them for good.
From our working notes on the audience
The answer isn’t a better crawler. It’s a better boundary.