# Tom Barfield, Director of Partnerships at Docket Legal (Chicago) -- read of LexiRisk, June 17, 2026

> 8 years evaluating legal workflow tools, currently helping mid-market law firms adopt contract automation. I have sat through more "AI will read your contracts" demos than I care to count.

## How I got here

Someone posted a link to Wishdeal Studio in the Legal Innovators Slack with a comment along the lines of "these guys are interesting, they actually publish their scoring." I clicked LexiRisk specifically because contract AI is my lane and I was curious what angle they were pitching. This was my Blue Line commute home, 6:08 PM, standing, one hand on the overhead rail, phone in the other. My daughter texted me twice during this read and I still kept going, which means something.

## What I clicked first

The specific list in the features section: "Non-competes, uncapped liability, auto-renewal traps, buried IP clauses." That earned my next 90 seconds. Not because it's clever, but because it's accurate. Those are the exact four things that bite founders who sign without a lawyer. The order even makes sense. Whoever wrote that list has actually read bad contracts, or at least talked to people who have.

But then I realized something that took me an embarrassingly long time to clock: this is not a LexiRisk product page. This is Wishdeal Studio selling me the *idea* of building LexiRisk. The hero has a "Try it Live" section and a "Before / With LexiRisk" comparison that reads like a real product demo. It is not. The product does not exist. I had to read "Honest disclosure: we don't have live customers on this idea yet" twice before I understood what I was actually looking at.

## Where I paused

"financial upside: 2/10"

They published that. On the page. For the product they are trying to sell. "$-13,500 Year-1 take-home (Fermi)" and "1 in 7 meaningful-success odds." The studio is telling you their own idea is financially weak before you hand them $99. That is either genuinely unusual honesty or a very sophisticated way to pre-empt the refund request. I re-read that section three times. I still am not sure which it is, but I respected it enough to keep reading.

## What I distrusted

"Our ML model catches the clauses that quietly ruin your day six months later."

There is no ML model. The product does not exist yet. That sentence is written for a thing someone else is supposed to build after buying the dossier. The tense is wrong for what the page actually is. The "Try it Live" section has a result interface mocked up as if you can upload a real contract right now. You cannot. That gap between implied and real is the part that bothered me most, not because it is deceptive exactly, but because it muddies the page's actual purpose.

Also: "Skeptic memos (12)" appears in the sidebar with no explanation. Are those memos from people who reviewed the idea and liked it? From people who said it was a bad idea? From competitors? No context at all. That number is doing nothing for me without a sentence explaining what I am clicking into.

## What would convince me

A single real output. Not a mock UI. An actual NDA or vendor MSA, run through whatever model exists, with real flagged clauses and real plain-English explanations. Even a rough one. Even a bad one. Something I can screenshot and send to a colleague and say "look what this thing caught." The proof that someone ran one real contract through one real tool would do more than everything else on this page combined.

The pain-intensity score of 10/10 is interesting but it is self-reported by the team that built the scorecard. I want to read those 12 skeptic memos. Not the headline, the actual content. That is the social proof I would trust here.

## What I'd ask in an email reply

1. The "Try it Live" feature in the hero: is there an actual working prototype I can drop a real contract into, or is that UI aspirational art for the person who buys the $99 tier to build?

2. When you write "Our ML model," what is actually underneath it? Fine-tuned model, GPT-4 with a structured prompt, something else? Not asking for the IP, just asking if the phrase means something specific or is shorthand for "AI generally."

3. Can I read the 12 skeptic memos before buying anything? That is actually the thing I want most from this page.

## Verdict: on-the-fence

The honesty about the financial weakness is rare enough that I trust these people are not trying to hustle me. But the page is trying to serve two audiences at once, the person who wants to use LexiRisk and the person who wants to build it, and it does not fully work for either. If a working prototype exists anywhere, even rough, I would reply.

---
*Memo by skeptic persona, generated 2026-06-17. Studio breaks own self-grading loop.*