Business users are about to ask far more questions of your data warehouse than budgets were ever built for.

On the podcast, TextQL CEO Ethan Ding explained the flywheel: once people can ask questions in plain English, the queries don’t just increase — they can explode. Someone starts with “what is this number?”, then “why is this number?”, then “given my role, what should I do next?”… and soon that question is running daily, then hourly. Multiply that across an entire company and your Snowflake/Databricks compute becomes a direct complement to your LLM token spend.

In Ethan’s words, it’s plausible to see 10,000× more queries hitting the warehouse — and every one burns compute.

So how do you avoid going bankrupt on analytics? I asked Ethan how TextQL keeps costs in check. The core idea: every query teaches the system, making the next query cheaper to run — so customers pay materially less per query six months in than on day one.

Every question has a price. Make sure the answer is worth it.

🎧 Listen

‍