# Jordan Hecht, Analytics Engineering Manager at Fendly (82-person B2B SaaS) — read of SemanticSQL, June 18 2026

> "9 years in data, currently managing a 3-person analytics eng team that spends half its time answering the same Slack questions from our CS and RevOps orgs. BigQuery + dbt + Looker stack, with some legacy Redshift we're slowly killing."

## How I got here

Googled "natural language SQL interface BigQuery teams" after my head of CS asked me for the fourth time this quarter to pull the same cohort retention table. A Reddit thread mentioned a few tools in this space and one commenter linked here. I had 12 minutes before standup.

## What I clicked first

"Query all your data in plain English" is exactly what I've been looking for a reason to dismiss. I've tried three of these. They all hallucinate joins on tables they don't understand and produce SQL that looks right until it quietly counts things twice. So the hero pulled me in because it named my pain and pushed me away simultaneously.

The spec table was actually the right call. Listing PostgreSQL, MySQL, BigQuery, Snowflake, Redshift, DuckDB in one row told me they at least know their market. Most competitors in this space pretend only Postgres exists.

## Where I paused

This section stopped me cold:

> "Honest disclosure: we don't have live customers on this idea yet. We shipped the strategy package; you ship the customer conversations."

I had to re-read that twice. I came here thinking this was a product I could sign up for. The "Get API key" button in the nav, the "Start Now (Free)" CTA, the "Results appear in milliseconds" copy all read like a live SaaS. Then I hit that block and realized I'm not looking at a product page. I'm looking at a Wishdeal Studio idea listing. The $5 unlock is for a dossier. The $99 is for "working code starter." They're selling a business plan, not software.

That's a legitimately interesting model but I almost missed it entirely. The framing is doing too much work trying to look like a product.

## What I distrusted

"Our semantic query engine understands context, table relationships, and business logic. It generates production-grade SQL automatically."

Every single tool in this category says something like this. The phrase "production-grade SQL" is doing a lot of lifting. What does that mean when your schema has 200 tables and 40 of them are deprecated and 6 share similar column names because three engineers named things without coordination over 5 years? That is the actual environment where this either works or doesn't. No mention of schema complexity, no mention of how it handles ambiguity, no example SQL output anywhere.

Also: "Reduce your database query costs by 40% through intelligent deduplication." Where did 40% come from? No footnote, no methodology, no range. That number smells made up or cherry-picked from one customer's edge case. Except there are no customers.

And the "67/100 Adoptability" self-score with "financial upside: 1/10" is either the most honest thing I've seen on a product page in years or a weird way to sell a business plan nobody asked for.

## What would convince me

A screen recording of someone on my team (not a developer, a CS ops person who writes occasional Salesforce reports) asking a genuinely ambiguous question like "show me accounts that churned in Q1 but came back in Q2" and watching the tool figure out which tables to join, handle a naming conflict, and produce SQL I could actually ship. Not a polished demo video. A real session with real confusion.

That would tell me more than any copy on this page.

## What I'd ask in an email reply

1. When you say "semantic query engine," is that a fine-tuned model, prompt engineering over GPT-4, or something proprietary? I'm asking because my legal team is going to ask where my query text and schema go.

2. How does it handle ambiguous column names across databases, specifically when two tables have a column called "status" that means completely different things in each context?

3. The page says this is a strategy package with no live customers yet. Is there an actual working prototype I can poke at, or would I be buying into something that's still in someone's Cursor session?

## Verdict: on-the-fence

The honest disclosure about having no live customers is genuinely rare and I respect it. But the page is doing too much identity confusion between "working SaaS product" and "business idea for sale" and I'm not sure which one I'm looking at even after reading it twice.

---
*Memo by skeptic persona, generated 2026-06-18. Studio breaks own self-grading loop.*