# Marcus Yee, Creative Director at Halftone Studio — read of Foreword, May 31 2026

> 9 years running a 6-person brand and web studio in Portland. We do maybe 40 client engagements a year. I drive a 2014 Subaru Outback and have a 7-year-old who is currently obsessed with Legos, which means I relate deeply to "scope creep by default."

## How I got here

A freelance developer I've collaborated with twice posted something on LinkedIn, something like "this saved me two revision rounds on a SaaS rebrand" with a screenshot of a brief PDF. No affiliate link, no promo code, just a screenshot. That's the kind of referral that actually makes me click. I googled "Foreword brief AI" and landed here directly. I've got maybe six minutes before my next call.

## What I clicked first

The headline pulled me: "Stop starting projects blind." That's the exact phrase I used in a retrospective with my team three months ago after a branding project went sideways. So I kept reading. The three-problem breakdown is accurate enough that I felt a little called out -- specifically the "disputes with no paper trail" one. I've been in that conversation. It's not fun.

What I actually clicked first was the sample output. I always go straight to examples because marketing copy on these tools is meaningless until I see what the artifact looks like. The Horizon Wellness brief is genuinely good. The "Not in scope" section under deliverables is doing real work -- that's the section that prevents the 11pm "can you also do social templates?" text.

## Where I paused

The intake form description: "Clients answer confidently without a PM holding their hand, and nothing gets skipped or filled in with 'TBD.'"

I want to believe that. But I've watched clients fill out Typeform intake forms I built myself, forms with clear labels and example answers, and they still wrote "TBD" in the timeline field and "not sure yet" in the budget field. So the question I'm sitting with is: what does Foreword actually do when a client gives garbage input? Does the AI flag it? Make something up? The page says "flags ambiguous answers for your review" but doesn't show me what that looks like. That would have been the screenshot to include.

## What I distrusted

"74% fewer revision rounds." I want to know where that comes from. Is it a survey of their 2,400 users? Is it one customer who was a statistical outlier? Is it a made-up benchmark? The number is specific enough to feel credible but there's no footnote, no "based on X responses from Y users in Z timeframe." That's a claim that deserves a source.

Also "2,400+ briefs generated" is a small number for a product making category-level claims. That's a few months of a mid-sized agency's output. It's not nothing, but it's not "we have enough data to make statistically meaningful claims about revision reduction" territory either.

The "AI handles 90 percent of the drafting" line is the kind of thing that sounds great until the 10% that requires my judgment turns out to be 40% on an ambiguous project.

## What would convince me

A video walkthrough of a real intake response that came in messy -- client said "ASAP" for timeline, left deliverables vague -- and showing exactly how Foreword flagged it and what the brief looked like before and after the creative director cleaned it up. I want to see the failure mode, not just the polished output.

Or: one named studio (not "a Portland agency", an actual name I can look up) saying "we used to have X disputes per quarter and now we have Y." Real numbers from a real business I can verify or at least Google.

The Horizon Wellness sample is good but it's clearly ideal-case input. Show me what happens when the client is the problem, which is most of the time.

## What I'd ask in an email reply

1. When a client leaves inline comments requesting scope changes during the approval step, does Foreword track that as a version or does it just overwrite the brief? I need to know if there's an audit trail on the negotiation, not just the final approved doc.
2. The white-label plan says intake links live on my domain -- is that a CNAME setup or do I need to mess with DNS records? My developer handles our infrastructure and I want to know if this is a 10-minute thing or a project.
3. What happens to my brief library if I cancel? Do I get an export? Is it locked?

## Verdict: curious-enough-to-reply

The sample output is genuinely the best one I've seen in this category -- it reads like something a senior PM wrote, not something a chatbot summarized. The pricing is low enough that the risk of trying it on one real project is minimal. The 74% stat still bugs me, but not enough to close the tab.

---
*Memo by skeptic persona, generated 2026-05-31. Studio breaks own self-grading loop.*