# Tom Szymanski, Owner / Creative Director at Keel Studio — read of Foreword (brief-ai), June 2, 2026

> 9 years running a 3-person web and brand studio in Chicago. Written probably 300 project briefs in my life. Had exactly three clients tell me "that's not what I said" six weeks deep into a fixed-price project.

## How I got here

LinkedIn ad. I almost always skip them but it said something about scope creep and I had literally spent that morning on a call with a client trying to add a fourth deliverable to a project we were two weeks into. I clicked. I've looked at maybe five tools in this space over the past two years. Most are glorified Google Form builders with a Notion export button.

## What I clicked first

"Stop starting projects blind" is a decent opening. Specific about the failure, not just the fix. Then I hit the three stats: "74% fewer revision rounds / 3 min average brief time / 2,400+ briefs generated." I read all three and immediately looked for a footnote. There isn't one. The 74% is doing the most work on this page and it has zero sourcing. I kept reading but that number sat in the back of my head the whole time.

Side note: the top of the page has "Unlock the dossier $5" and "See the build below" which made no sense to me on first pass. I thought I'd landed on the wrong page. Took me a second to figure out this was some kind of product showcase wrapper around the actual product page. That's friction I didn't expect.

## Where I paused

The Horizon Wellness sample brief. Specifically this line: "Deadline is firm -- client has a lease signing event June 20 and requires assets for the venue signage vendor by that date." That detail. That's the kind of thing that normally lives only in my head or buried in a Slack thread from two months ago and then bites me when I've already filed the project as done. If the intake form surfaces that and the AI locks it into the brief, that is genuinely useful. I scrolled back up after reading that section.

## What I distrusted

The 74% number, still, no source. "2,400+ briefs generated" is actually a small number if this product has been running for any real stretch of time. It's not reassuring social proof, it's the kind of number that makes me wonder how new this is and whether the AI has seen enough real-world project data or if it's just a GPT wrapper around a brief template. Also "The AI handles 90 percent of the drafting so you can focus on the 10 percent that requires your judgment" is a tidy line that I've heard some version of from every AI tool I've looked at in the last two years. Means nothing without seeing what the 90 percent actually looks like when the client has filled out the intake form badly.

## What would convince me

One named studio, with a domain I can look up, saying what their revision round average was before and after. Not a percentage pulled from nowhere, an actual account. "We ran 40 branding projects last year, went from 3.1 average revision rounds to 1.3." That would land. Also: show me what happens when a client types "something modern but also classic, you know?" as their brand direction. Does the system flag it as ambiguous and tell me to follow up, or does it launder that vagueness into polished brief language that still means nothing?

## What I'd ask in an email reply

1. Where does the 74% come from? Survey, internal platform data, controlled comparison? I'll ask because if the answer is "our beta users estimated it" I'm going to recalibrate how much I trust the rest of the claims.
2. The "flags ambiguous answers for your review" in step two -- can I see a screenshot or a real example of what a flagged brief looks like? That's the feature I'd actually pay for and I can't evaluate it from a description.
3. White-label on the Studio plan: does "Intake links live on your domain" require me to set up DNS records or is it a subdomain redirect you handle? I have clients who notice URLs and I don't want to explain why their intake form is at foreword.app/my-studio.

## Verdict: on-the-fence

The sample brief is the best thing on this page and it almost gets me. But I'm not handing over $79/month based on an unsourced stat and no named customers. One real case study and a straight answer on the 74% and I'd probably start a trial.

---
*Memo by skeptic persona, generated 2026-06-02. Studio breaks own self-grading loop.*