Making Ordinary Companies AI-Ready
Leah Weiss, co-founder of Preql, on delivering clean data to LLMs
The core mistake of the modern data stack era was selling Fortune 50 data ambition to companies that barely had a data team. For most businesses, the hard part was never buying Snowflake, dbt, or BI tools. It was hiring people who could keep pipelines running, define metrics the same way across teams, and turn that work into revenue or better decisions. That mismatch is the opening Preql is built around.
-
At WeWork and later in consulting, Leah Weiss saw that ROI broke not because tools were modular, but because every layer needed specialized operators. Small and mid sized companies were sold the same playbook as companies with huge internal data orgs, and the operating burden swamped the payoff.
-
A close comparable is Equals, which won downmarket finance teams by avoiding the full warehouse plus ETL plus transformation plus BI buildout. Its pitch was simple, work across Stripe, HubSpot, Postgres, and spreadsheets without a six month implementation and a new analytics hire.
-
This is why Preql focuses on the messy middle, data spread across systems, business logic trapped in Excel, and no single trusted metric definition. Instead of assuming a clean warehouse, it tries to replace manual cleanup and semantic translation work that most companies never staff well enough to do themselves.
The next phase of the market shifts from building elite internal data teams to buying software that acts like one. As AI makes every company want chat, copilots, and eventually automated decisions on top of internal data, the winners will be the vendors that make ordinary companies usable by AI without requiring them to become Facebook or Google first.